1

7/27/2014 Draft Please Do Not Quote or Cite

AN AGENCY THEORY OF PATENT LAW:

LINKING INNOVATORS AND INVENTION USERS

by

Richard S. Gruner*

Copyright 2014

ABSTRACT

This article presents a reinterpretation of patent laws as shapers of agency relationships between

innovators and innovation users.1 Patent rights reflect latent incentives encouraging potential innovators

to understand, analyze, and solve the practical problems of innovation users. In this way, patent rights

link and align the interests of innovators and innovation users. They make innovators specialized agents

of innovation users in the pursuit of technologically distinctive and patentable innovations.

By presenting the promise of rewards for serving the interests of innovation users, patent rights

both attract the attention of technology specialists to the solution of practical problems they might

otherwise overlook and press these specialists to explore the boundaries of their technical fields to

propose outlier solutions that will meet patent law tests for intellectual property rights. The results are

both more attempts at technical solutions and more breadth in the content of those solutions. In these

ways, patent rights serve fundamental roles in incentivizing technical advances and improving the useful

arts, all without the transaction costs of contract formation efforts typically underlying agency processes.

Patent rights are, in short, the linchpins to an ongoing system inviting and luring technical specialists to

* Richard Gruner is a Professor Emeritus at the John Marshall Law School in Chicago and former Director of

the Center for Intellectual Property Law there. He is a registered patent attorney and a former inside

counsel for the IBM Corporation. Professor Gruner is a member of the New York and California state bars

and a graduate of the University of California, Irvine (PhD, Criminology, Law and Society 2008),

Columbia University School of Law (LL.M. 1982), the USC Law School (J.D. 1978), and Caltech (B.S.

1975). He is the coauthor of Transactional Intellectual Property: From Startups to Public Companies

(Lexis/Nexis 2012) and Intellectual Property: Private Rights, the Public Interest, and the Regulation of

Creative Activity (West 2nd ed. 2010).

1 The term "agency relationship" is used here in a behavioral sense, not as a reference to a legally recognized

"agency" as defined by "agency law." Agency law primarily defines the powers and liabilities created

when two persons enter into an agreement that one of them (the agent) will act on behalf of the other (the

principal). By contrast, the term "agency relationship" is used here to refer to the conduct of one person

seeking to promote the interests of another regardless of whether this course of action results from a formal

contract or from other structured motivations such as the promise of valuable patent rights upon the

completion of certain successful innovation efforts.

2

serve as agents of innovation users, with the acceptance of the ongoing invitations manifested by

successful completion and commercialization of widely useful inventions.

I. Introduction

This article describes how patents serve as substitutes for privately negotiated contract terms in

defining incentives for potential innovators with specialized technical knowledge or practical experience

and linking the interests of those innovators to the interests of potential innovation users. Patents not

only align the interest of these parties by giving innovators a potential stake in solving the practical

problems of innovation users, they serve to align and scale the concerns and resource allocation of

inventors to the scope and depth of demand for inventions by potential users. The rewards available

through patent rights incentivize inventors to look to the needs of invention users for definitions of useful

advances. The size of these rewards are scaled (at least roughly) to the degree of utility achieved for

invention users via successful invention projects, meaning that big rewards attach to highly useful (or at

least popular) technical solutions and lesser rewards to less useful advances. This, in turn, encourages

attention to projects and the use of research resources in proportion to the utility of solutions being

pursued. In both attracting the efforts of innovators and signaling the importance of innovation projects,

patent rights encourage inventors to act as agents of invention users in efficient ways. And the formation

of these agency relations is achieved without the need for innovation users to identify and form agency

contracts with potential innovators. Consequently, patents overcome many of the informational and

transactional limitations of private agency contracts as means to encourage innovation progress.

The agency-focused interpretation of patent law advocated here has several advantages as a

means to understand the impacts of present patent laws and to project needed reforms. First, it

emphasizes the technology-enhancing goals of patent laws over sometimes misleading considerations of

property rights and controls.2 Patent law standards are desirable under this model solely if they promote

the interests of principals in acquiring new technologies. Rights and incentives provided to innovators are

viewed as instrumental means to promote the interests of innovation users and consumers who occupy the

position of principals in the patent-structured relationships of interest.

Second, by viewing innovation efforts as agency processes and assessing patent rights as one

feature shaping these agency processes, previously developed methodologies for analyzing agency

processes can be applied to patent-mediated innovation processes, leading to expanded descriptive and

normative studies of the patent system. Descriptive studies employing an agency model can help us

assess the probable impact of present patent standards in promoting innovations of benefit to invention

users. Normative studies using this model can suggest how patent incentives should be adjusted to bring

innovative efforts serving invention users to more efficient levels.

Third, by viewing patent rights as tools for shaping agency relationships promoting innovation we

can gain new insights about situations where patent rights should and should not apply. The likely

2 The model proposed here recognizes that property controls over innovations may provide means to reward

innovators for the successful pursuit of the interests of user-principals in innovative agency relationships.

Property rights concerning patented inventions are treated in this model as a form of compensation awarded

to innovators who have met their part of the innovation bargain implicit in patent laws by diagnosing and

solving some practical problem of potential invention users. The key questions governing the proper scope

of patent rights within this model are what sorts of rights will properly align the interests of innovation

developers and users and under what circumstances should the property rights flowing from the issuance of

a patent be recognized under our legal system?

3

success of innovation-focused agency processes dictates the appropriate boundaries of the patent system

under this approach. In settings where agency processes are unlikely to be effective in promoting useful

innovation, associated patent rights will also serve no useful role and may imposed wasteful costs or

restrictions on business activities. Seen this way, the scope of potentially effective agency processes for

innovation helps to delineate the proper boundaries of the patent system and provides a new means to

study the proper scope of patentable subject matter.

The analyses in this article proceed in five parts. Section II describes the proposed agency model

of patent rights in which patents are components of broader patent-mediated agency processes for

technological innovation. The model is developed by initially considering the agency features of

innovation projects involving one innovator working on behalf of one potential innovation user. This

version of innovation involves a single principal (the potential innovation user) and a single agent (the

innovator). Many of the agency features present in this simple project are also present in larger, multi-

agent projects but can be studied more effectively and clearly in the simpler type of project. Moving from

this simple case, I then examine extensions of the model along two dimensions. One extension considers

the agency implications of including multiple potential innovators as competing parties seeking to

produce an advance for a single potential user. This extension involves a single principal (the potential

invention user) working with multiple agents (the competing innovators). The second extension

examines the agency implications of moving from a single potential user to a setting where potential

innovation users are numerous and diverse, but nonetheless share a common interest in a particular

functional innovation feature that makes the group of users a common target of a single innovation

project. This type of innovation involves multiple principals (comprised of the members of the user

group) being served by multiple agents (the multiple competing innovators).

Part III summarizes agency theory and suggests how it may be applied to interpret patent-

mediated agency relationships. Features and limitations of patent incentives as means for shaping and

administering innovation-focused agency processes are also explored.

Part IV uses the proposed model to identify patent law features that affect the impact of patent

rights on innovation-focused agency processes. This discussion will critique a number of present patent

law doctrines in terms of whether they successfully align the interests of invention users and innovators

and thereby create desirable agency relationships.

Part V considers the outer boundaries of the proposed model and implications of these boundaries

regarding the proper scope of patentable subject matter. I argue that patent incentives are only needed to

promote agency processes for innovations with substantial user sets where direct contracting processes for

encouraging innovation are unlikely to be effective. This will be the case where numerous potential users

can benefit from an innovation (through either widespread use or repeated use of the innovation or both)

and the innovation can be described and transferred to users in a systematic manner without direct

interaction between the innovator and the users of her innovation.

Finally, in Part VI I consider several features of the patent system that may deserve

reconsideration in light of their impacts on agency processes for innovation. The breadth of the patent

law characteristics considered -- including such diverse patent law features as pre-issuance patent

application publication, non-obviousness standards, and rewards to patent applicants in corporate

environments -- is indicative of the wide-ranging value of agency theory in critiquing present patent

standards and shaping future patent policy.

II. Components of an Agency Theory of Patent Law

4

The agency theory of patent law described here treats patent rights incentivizing tools to activate

actions by technology specialists working as agents on behalf of innovation users. Patent rights will

incentivize innovation and support agency relationships between inventors and invention users if those

rights function as substitutes for directly negotiated innovation payment terms negotiated between

innovators and innovation users. In actual patent-mediated innovation processes, there are no such

negotiations. Innovators self-identify themselves as possible inventors and project the potential needs of

innovation users. They then act as agents of innovation users by seeking successful inventions despite the

fact that no contract or other private arrangement for payment for inventive labor exists between the

innovator and the parties who will benefit from innovation efforts if they are successful. The payoffs

implied by patent rights provide the substitutes for these missing agency contract terms. They both

encourage potential innovators to consider acting as agents for innovation users and establish the

mechanism for agency payments once the needs of principals (that is, innovation users) are served by

successful inventions.

This section provides a summary of this agency framework for interpreting patent rights.

A. Establishing a Context and Purpose for Patent Rewards

Patent laws and related restrictions on patented inventions are traditionally justified as means to

create rewards for inventors that promote socially beneficial innovation.3 Patents are intended to

“promote the Progress of … [the] useful arts” by increasing specialized knowledge about useful items and

processes and applying that knowledge to produce more such items and processes4 Patent rights afford an

inventor exclusive control over such commercially significant activities as the making, using, selling, and

importing of a patented invention to create exclusive commercial opportunities and gains that reward the

inventor for his or her inventive efforts.5 The scope of the potential rewards varies with the value and

popularity of the resulting invention and the willingness of users to pay substantial patent-influenced

prices for the invention.

The promise of patent-influenced rewards is aimed at encouraging persons with specialized

capabilities or insights to apply their efforts to developing useful innovations. John Stuart Mill captured

the essential reward logic underlying patent rights with his observation that an inventor producing a

socially valuable advance "ought to be both compensated and rewarded" through the grant of a temporary

"exclusive privilege" conveyed by a patent.6 Mill felt such a patent-based reward scheme was preferable

to other innovation incentivizing schemes -- such as government-issued prizes or rewards for useful

advances. Patent-influence rewards delivered via commercial transactions and processes were superior in

Mill’s estimation because such rewards avoided any need for discretionary action by parties authorizing

rewards, secured commercial rewards to inventors in amounts varying with the value and usefulness of

inventions as seen by users, and ensured that the rewards given inventors were ultimately paid for (via

3 See, e.g., Mark F. Grady & Jay I. Alexander, Patent Law and Rent Dissipation, 78 Va. L. Rev. 305, 310

(1992) (noting the “long intellectual history of reward theory" in justifying patent rights).

4 United States Constitution, Art. I, Sec. 8, Clause 8.

5 See, e.g., Mark F. Grady & Jay I. Alexander, Patent Law and Rent Dissipation, 78 Va. L. Rev. 305, 310

(1992).

6 2 John Stuart Mill, Principles of Political Economy 548 (D. Appleton & Co. 1901) (London, 5th ed. 1862).

5

background processes resulting in patent-elevated invention prices7) by the users who benefited from the

inventions.8

This view provides a framework for thinking about both patent rights and the types of patent

rewards that will best serve the aims of the patent system. If patent rights are created and administered to

incentivize innovation on behalf of invention users, then these rewards will be justified up to (but not

beyond) the point where they attract capable innovators to undertake innovation projects that are cost-

effective taking into account their likelihood of success, the social value of the innovations they target, the

costs of completion, and the further opportunity costs of the alternative activities by innovators that

inventive efforts displace (that is the costs of the foregone activities that innovators cannot complete

because they are busy with innovation projects).

Mill’s framework treats patents as instrumental means to economize on the time and talents of

persons who are capable of producing patentable advances.9 In order to encourage these persons to

allocate their time to work on the most socially beneficial types of advances, it is desirable to link the

efforts of potential innovators to the interests and invention gains of innovation users. This linkage will

“price” the time of talented individuals in accordance with the scope of public benefits that are likely to

result from various innovation efforts.10

Furthermore, in seeking these important advances, patent rights

encourage innovators to give special attention to designs that incorporate previously unproven and “non-

obvious” features of the sort needed to qualify for patent protections.11

7 Retail prices charged to users of patent-protected products or services tend to support large rewards to

inventors even if the inventors have transferred their rights to the relevant invention to their employer or

another commercial entity such that a party other than the inventor is the immediate seller of the patented

product. The terms on which transfers of the inventors’ rights are accomplished are influenced by the

transferee’s anticipation of patent-influenced retail sales revenues in the future. Hence, the salary and

bonuses given an employee working on patentable advances are influenced by his or her employer’s

knowledge that resulting patented inventions can be sold at patent-influenced prices in the future.

Likewise, the amounts that individual inventors who are patent holders are paid for assignments of their

inventions are influenced by the assignee’s expectation that the transferred patent will allow the assignee to

charge patent-influenced prices for related products or services in the future.

8 Id.

9 Under this view, the impact of patents must be judged not in light of how they are enforced, but rather at an

earlier point when the promise of enforcement influences the conduct of a potential innovator and others

(such as employers and resource providers) who interact with the innovator. The scope of rights upon

patent enforcement will bear on what value innovators (or their employers and resource providers)

associate with alternative projects as they contemplate directions for innovative efforts. Patents influence

choices about these directions. The degree to which they influence these choices will depend in part on the

strength of patent rights and the resulting projected value of anticipated patent enforcement patterns.

10 I have explored this innovation behavior pricing and prioritizing logic underlying patent law elsewhere.

See generally Richard Gruner, Dispelling the Myth of Patents as Non-Rivalrous Property: Patents as Tools

for Allocating Scarce Labor and Resources, 13 COLUMBIA SCI. & TECH. L. REV. 1 (2012).

11 Patents are limited to useful inventions that are both new and significantly different from prior advances in

the same field. See 35 U.S.C. § 103. This later requirement – imposed via tests requiring inventions to be

“non-obvious” to average practitioners in the relevant field – are applied primarily to disallow patent

rewards (and avoid patent system costs) for advances that reflect minor technical adjustments to prior

designs in the same field in which the new design is so similar to prior designs (or is made along such

familiar lines of previously successful adjustment) as to be predictably successful to many parties in the

field. Advances involving such minor adjustments are thought to be within the capabilities of many parties

and thus likely to result without the special incentives of patent rewards. By contrast, the sorts of non-

6

This view of patents solves a long-standing conceptual problem in patent law regarding the policy

basis for granting inventors exclusive “property rights” over the use of patented inventions. Such rights

are not needed to avoid exhaustion or inefficient consumption of the inventions themselves since these

inventions are ideas that cannot be exhausted. But patent rights resembling property rights can be useful

means to govern the use (and misuse) of inventors’ time and resources, which certainly are exhaustible

resources. The potentially scarce time and abilities of talented innovators need to be carefully allocated

via properly crafted incentives that both attract these parties to projects of large public interest and

prioritize the scope of innovators’ efforts among projects of public concern. Patents create and scale the

proper incentives by “pricing” the time and resource exhaustion of innovators in accordance with the

socially beneficial results they achieve (as mirrored in the commercial value of the patent-protected

products they enable).12

Hence, a property logic underlying patent rights is sensible because property

controls over access to patent-protected inventions activate a system economizing on the scarce resources

of inventors’ time. Property-based patent rights are not justified because inventions are scarce resources

but rather because the efforts and resources of the inventors who produce the inventions are scarce.

Property-based patent rights are simply part of a labor incentivizing and economizing system influencing

innovators via rewards gained from the products of their labor.13

Patent-influenced controls and rewards

obvious advances that patents encourage involve previously untried design approaches in the relevant fields

which will be rare and worthy of the special encouragement of patent rewards.

12 Robert Merges has recognized that intellectual property rights can be a means to “propertize” the labor of

an innovator and thereby reallocate how that labor is integrated into business enterprises. He has analyzed

the implications of intellectual property rights in the context of the work of a consultant:

A consultant generally can only sell a given unit of labor once, and she can sell it only to a single

firm. Intellectual property, however, in effect ‘propertizes’ her labor, making it possible to sell the

same unit of output multiple times to multiple firms. Of course, for this to work, the consultant

must produce something that intellectual property law protects, and she must retain ownership of

her work product, typically by contract. Assuming ownership of a protected work, however,

intellectual property rights allow her to transform her efforts from a onetime service into a

multiple-use commodity. This conversion of services into an asset that the producer can trade

many times of course enhances the potential economic returns from such work.

Robert P. Merges, Intellectual Property and the Costs of Commercial Exchange: A Review Essay,

93 MICH. L. REV. 1570, 1575 (1995). Merges argues that this propertization process will tend to encourage

innovators to pursue valuable advances for the advances= own sake, with confidence that they can gain

commercial returns on the resulting advances through specialized businesses engaged in developing

advances and licensing rights to use the advances.

While these insights into the impact of intellectual property rights on specialization and enterprise

organization dynamics are important, Merges does not address the further implications of the

“propertization” effect of intellectual property rights examined in this article. As described here, the

“propertization” of innovative labor achieved by attaching patent rights and associated rewards to the

production of patentable inventions may aid in economizing the allocation of labor and other resources

within a business or within the work choices of an individual and, hence, encourage the allocation of

additional innovative efforts towards activities and advances of significant social value.

13 Actually, the view of intellectual property rights advocated here focuses on property controls of innovators

over the fruits of their labors rather than over the labors themselves, with the controls over the products of

labors influencing what labors are undertaken in the first place. Robert Merges has noted this distinction in

a somewhat different context, giving the following explaining his views of the “propertization” of labors of

creative persons through associated intellectual property rights:

7

ensure that the scarce resources of talented innovators (and the also scarce resources of money and

equipment dedicated to innovation by companies14

or other organizations such as major universities that

support large-scale innovation) are not misallocated.15

Patents are instrumental means of attaching

legally-promised rewards and payments to choices by innovators to pursue socially beneficial innovation

activities.

B. Agency Relationships Resulting from Patent Linkage of Innovators to Invention Users

1. Characterizing Agency Relationships

Commentators at least since Locke have asserted that everyone owns his or her labor; in Locke's

case, this was an outgrowth of his starting point that everyone owns his or her own body. John

Locke, Two Treatises of Government 328-29 (Peter Laslett ed., 1960). But when these

commentators refer to a property right in one's labor, they are talking about the right to bargain for

a wage before engaging in work -- in essence, the right not to be a slave. By contrast, I am

referring to the conversion of labor into a tradeable asset or property right. "Assetization" might

be a more appropriate term for what I have in mind; but since this sounds even worse than

"propertization," I will stick with the latter.

Id. at 1575 n.12.

14 Corporations, even more than individuals, structure their resource allocations in accordance with the profit

potential of various alternative actions. Corporate decision-making processes typically force parties

requesting allocations of limited corporate requests to justify their requests in terms of the profit potential

associated with their requested use of the resources. In responding to competing resource requests

supported by this type of profit potential information, internal corporate resource allocation processes often

serve as internal capital markets for the allocation of corporate “investments” of budgets and other

corporate resources to projects in light of the projected profit potential of the projects. In these internal

capital market systems, potential invention efforts compete for resources with requests for many non-

innovative projects that are also potential sources of increased corporate profits (such as additional

marketing efforts for old products or additional cost cutting measures). Whichever actions seem likely to

produce the greatest corporate profit returns will tend to be seen as the best use of corporate “capital”

reflected in available corporate resources and will tend to receive those resources as company managers

make commitments to certain projects and reject others. See R. Gruner, S. Ghosh, & J. Kesan,

TRANSACTIONAL INTELLECTUAL PROPERTY: FROM STARTUPS TO PUBLIC COMPANIES (Lexis/Nexis 2012);

George G. Triantis, Organizations as Internal Capital Markets: The Legal Boundaries of Firms, Collateral,

and Trusts in Commercial and Charitable Enterprises, 117 Harv. L. Rev. 1102, 1105, 1109-15 (2004).

15 In this regard, patents are not so much about avoiding overconsumption of a resource -- the so called

“tragedy of the commons” -- as they are means to avoid an inverse “tragedy of misallocation” that can

result where resources are not allocated to their highest use -- that is, the most socially valuable use. The

asset that is at risk of misallocation absent patent rights is the labor of those highly skilled individuals who

are most likely to produce patentable innovations. Absent patent rights, the labor of these parties may be

allocated to tasks other than the sorts of unusual advances with high social value that these individuals (and

perhaps only these individuals) are capable of producing. The avoidance of this type of misallocation is,

according to Harold Demsetz, one of the primary justifications for recognizing property rights is that, in a

world of no transaction costs, such rights will internalize the costs and benefits of an activity in the rights

owner. See Harold Demsetz, Toward a Theory of Property Rights, 57 AM. ECON. REV. PAPERS & PROC.

347, 348, 349 (1967). By attaching property rights in the form of patent rights to certain types of unusual,

non-obvious innovations of talented individuals and connecting rewards to the patent rights that are equal

to the net social benefits of patented inventions, patent laws can aid innovators in internalizing the positive

externalities associated with decisions about whether to pursue innovative activities and thereby improve

the quality of those decisions in efficiently allocating the labor of talented individuals.

8

Agency relationships for innovation are all around us, accounting for many of the innovative

products and expert services in our complex society. Agency relationships involve actions taken by one

person (the agent) on behalf of or for the benefit of another (the principal).16

Typically, an agent will

receive a reward for successful completion of the tasks desired by a principal so as to align the incentives

and interests of the agent with those of the principal. However, flaws in information gathering and other

transactional processes may frustrate the implementation of ideal incentives for agents and lead to the

imperfect alignment of principal and agent interests. Principals may imperfectly monitor the completion

by agents of desired actions, leading to the granting of rewards to agents when no benefits for principals

have actually been achieved. Principles may also pay too little or too much to agents for the completion

of tasks desired by the principals, thereby causing the agents to devote too little or too much effort to the

completion of the tasks given the value of the tasks to the principals. These types of errors lead to

significant gaps and inefficiencies in the links between rewards to agents and benefits to principals. Such

gaps can cause agents to diverge significantly from the efficient attainment of principals' goals.17

These types of agency processes and problems have been the focus of extensive studies by

economists and others. Most studies of agency relationships and processes have focused on agency

arrangements created in direct transactions between principals and agents resulting in contractual

relationships between the parties. The resulting contracts typically define contingent rewards and

incentives aimed at encouraging agents to take specific actions serving principals’ interests. For example,

a typical, contract-based agency relationship arises where a homeowner engages a neighbor’s child to

mow the homeowner’s lawn for $10.00, with payment to occur when the lawn is completely mowed. If

all goes well, the child (acting as the agent) will mow the lawn and serve the interests of the homeowner

(acting as the principal). The withholding of payment until completion of the mowing both serves to

make this promised payment an incentive for diligent work by the child in completing the mowing and

also gives that homeowner a chance to monitor the results and check the quality of the agent’s efforts

before delivering the promised reward (thereby providing some assurance that the reward is not being

paid for no effort or flawed efforts). This simple agency transaction incorporates all of the major features

of an agency relationship and suggests how commonplace agency relationships are in our everyday

activities.

However, economists have recognized that face-to-face transactions leading to contract-based

agency relationships are not necessary to create agency incentives and processes. Agency arrangements

resulting from contracts are merely a familiar and relatively simple subset of all agency relationships.

The full range of agency relationships includes all arrangements in which one party’s actions serve the

interests of another, usually in response to contingent rewards or penalties encouraging such actions.

Recognizing that agency processes go far beyond contractual relationships, Steven Shavell has

described the breadth of agency processes as follows:

In [agency relationships], one party, the principal, “enjoys” the outcome of the activity of the

other party, the agent. The agent's effort (or expenditure or, more generally, his action) together

with a random element determines the outcome. The principal then pays the agent a fee. For the

case of a professional and his client this description of the principal and agent relationship is

obviously appropriate. The description may be seen to apply to [other cases such as the

relationship between insurer and insured, shareholders and management, and even society and a

16

See generally K. Eisenhardt, Agency theory: An assessment and review. 14 ACADEMY OF MANAGEMENT

REVIEW 57, 57-74 (1989).

17 See id.

9

polluting firm] and, indeed, to any relationship where only one of the parties directly influences

the probability distribution of the outcome.18

In these settings, the value of the outcome to the principal or the "fee" paid to the agent can be positive or

negative. For example, in a principal-agent relationship between society and a polluting firm, both the

outcome received by the principal (pollution costs suffered by society) and the "fee" applied to the agent

(the costs in fines or other penalties imposed on a polluting firm) are negative.19

The simplest forms of

principal-agent relationship will involve promises of positive payments to an agent for the realization of

positive results for a principal (such as the lawn mowing relationship already discussed). However, an

entirely parallel – and economically equivalent – form of agency relationship can be structured in which

negative consequences are threatened for an agent who produces negative results for a principal. The

potential advantages (and flaws) of such a negatively-framed agency relationship are equivalent to those

in the positive form of the relationship. For simplicity sake, most of the discussions of agency

relationships in this article with address relationships involving positive performance results achieved

through positive rewards to agents.

In the innovation situations analyzed here, agency relationships influenced by patent rights

encourage innovators to examine the problems and needs of potential innovation users and to design

technological solutions that provide practical benefits to those users. In both diagnosing the practical

problems of potential invention users and in designing solutions, inventors serve as agents of the users

they benefit.

Patent rights (and the commercial opportunities they imply for popular inventions) encourage

inventors with special knowledge or abilities to look beyond themselves and their own circumstances and

consider how technology might be used to serve others. Inventive efforts frequently produce innovations

that benefit users who are very different from the innovators themselves (who, at least in today’s complex

innovation environments, are often highly trained and specialized masters of narrow areas of knowledge

and technology).20

The promise of rewards based on patent rights supplies innovators acting as agents of

the potential users of their advances with important incentives to look to the needs of those users and

shapes the focus and scope of their innovation efforts. The promised "payoff" enabled – but not fully

18

Steven Shavell, Risk Sharing and Incentives in the Principal and Agent Relationship, 10 Bell J. of Econ.

55, 55 (1979).

19 See id.

20 Of course, an innovator may create an advance to aid him or herself, perhaps because the innovator

operates or is part of a business that can use the advance to later advantage. In such circumstances, a party

may just keep the advance as a secret within that business rather than share it with competitors in ways that

will aid the competitors. In such settings, patent rewards still serve a purpose in enhancing the likelihood

that an innovator will disclose and commercialize an advance as an agent of potential further users beyond

the innovator.

Furthermore, even where an innovator has some personal motivation to pursue an advance for

self-benefit, the perspective-enhancing impacts of potential patent rewards may change the type of

innovation the individual pursues. The needs and interests of the bulk of users of the advance may be

different (or broader) than those of the original innovator. By looking to the full set of potential users as

sources of potential rewards – as patents will tend to promote – an innovator is encouraged to adopt the

design perspective and goals of the broader user set and to use the needs of that broader set of users as the

measure of a successful invention. Thus, even where an inventor is a potential user of an advance, his or

her inventive efforts may be valuably influenced and improved by the perspective-setting agency influences

of promised patent rights.

10

guaranteed – by patent rights is the special "fee" paid to innovators for playing their parts in the patent-

mediated principal and agent relationships between invention producers and users.

Within this framework, various substantive and procedural criteria for recognizing and

administering patent rights may be better or worse in linking the interests of inventors and invention users

and in establishing desirable agency relationships for innovation. A descriptive assessment of present

agency characteristics and inefficiencies embedded in patent laws and procedures is contained in Section

IV of this article. Using the same agency perspective, a normative evaluation of how patent reforms

might improve the operation of innovative agency processes is contained in Section V of this article.

B. A Preliminary Distinction Between Agency Law and Agency Relationships

In the discussions which follow, it will be useful to remember that, despite the similarity of its

name, agency law has little to do with the agency relationships considered here.21

The agency

relationships described here are behavioral relationships not relationships defined or regulated by agency

law. 22

Agency law primarily addresses agency relationships created in face-to-face transactions, usually

involving explicit (albeit sometimes brief) agency contracts between the parties. 23

Agency law

supplements the terms of these contracts and describes certain consequences flowing from the

formulation of agency arrangements. This body of law is intended 1) to protect the interests of parties

affected by agency arrangements who are not parties to the agency contracts and 2) to specify certain

default terms defining key agency relationship features where the parties' own agency contracts are

incomplete and do not address these key features.24

21

The most recent draft of the Restatement (Third) of Agency describes the distinction between the

behavioral and legal definitions of agency relationships as follows:

"[T]he terminology of agency is widely used in commercial settings and academic literature to

characterize relationships that are not necessarily encompassed by the legal definition of agency.

... In economics, definitions of principal-agent relations encompass relationships in which one

person's effort will benefit another or in which collaborative effort is required.... Not all such

situations, however, amount to relations of agency within the legal definition.

Restatement (Third) of Agency ' 1.01 cmt. b (Tentative Draft No. 2, 2001). One key difference between

relationships covered by the non-legal and legal definitions of agency is that the latter involves

relationships in which the principal is capable of controlling the activities of the agent. See id. at ' 1.01 &

cmt. c.

22 Cf. Robert H. Sitkoff, An Agency Costs Theory of Trust Law, 89 Cornell L. Rev. 621, 636 (2004)(analysis

of trust law "[u]sing the vocabulary of agency in economic rather than legal parlance").

23 Agency law defines certain legal powers and potential liabilities created when two or more persons enter

into an agreement that one of them (the agent) will act on behalf of the other (the principal). In modern

settings where many complex tasks are accomplished by groups of coordinated actors, many important

societal functions are accomplished by agents acting in part under the governance of agency law. For

example, agency law covers such common relationships as those between employer and employee,

corporation and officer, client and lawyer, and partnership and general partner. See Restatement (Third) of

Agency ' 1.01 cmt. c (Tentative Draft No. 2, 2001).

24 Agency law defines a variety of relationships between principals and agents and between those parties and

outsiders:

"As defined by the common law, the concept of agency posits a consensual relationship in which

one person, to one degree or another or respect or another, acts as a representative of or otherwise

acts on behalf of another person with power to affect the legal rights and duties of the other

11

However, since it presumes some direct interaction and (generally) contracting between principal

and agent, agency law is not sufficient to cover innovation processes in which innovators lacking prior

contacts with technology users pursue the interests of those users as agents in creating useful advances.

In these circumstances, the agency relationships between innovators and innovation users must be defined

by means other than contract terms and related agency law provisions. As described in this article, the

applicable agency features are defined in part via patent rights, which substitute for some of the contract

terms and agency law provisions that might have resulted had innovator and innovation user dealt with

each other directly. Hence, while patent laws and resulting rights serve functions parallel to directly

negotiated agency contracts and resulting agency law consequences, patent-mediated agency relationships

of the types described in this article do not depend on either contract or agency laws.

C. An Analytic Starting Point for Understanding Innovation-Focused Agency Relationships:

A Single Principle-Single Agent Innovation Contract

Because they serve similar functions in promoting innovation and are somewhat easier to

understand, contract-based agency relationships targeting innovation offer a useful introduction and

comparative introduction to patent-based agency relationships. This subsection describes a simple form

of contract-based agency relationship aimed at promoting innovation. This simple relationship will be

used to illustrate some of the basic features of agency processes promoting innovation. The same agency

features appear in more complex forms in patent-based agency processes promoting innovation.

The simplest form of contract-based agency arrangement for innovation is a single principal-

single agent contract calling for a person with technical expertise (the agent) to create a new item serving

some need of an intended user (the principal). A contract for this type of effort is essentially a custom

product design and production contract calling for work “on spec” – that is, for the creation of an item

meeting certain design specifications dictated by the acquiring party. The designer is expected to act as

the agent of the acquiring party in the course of specifying design criteria and producing a workable

overall design. The acquiring party will, if a successful product is produced, receive the resulting item

(and perhaps control the rights to produce further units of that item). As the benefitted party, the

acquiring party will be the principal benefitting from that agent’s design efforts. The fee promised to be

paid to the designer upon successful completion of the contract creates incentives encouraging the

designer to keep the interests of the acquirer in mind while carrying out the contract and aligns the

interests of principal and agent in this context.

This type of simple agency arrangement includes three essential features found in all agency

relationships for innovation: 1) a definition of the actions or results desired from the agent pursuing

innovation, 2) the means and criteria the principal will use to evaluate the efforts or results of the agent as

a basis for authorizing payments to the agent, and 3) and the amount of the fees or other rewards the agent

will receive for successful performance.25

person. The person represented has a right to control the actions of the agent. Agency thus entails

inward-looking consequences, operative as between the agent and the principal, as well as

outward-looking consequences, operative as among the agent, the principal, and third parties with

whom the agent interacts."

Restatement (Third) of Agency § 1.01 cmt. c (Tentative Draft No. 2, 2001).

25 See Kenneth J. Arrow, The Economics of Agency in Principals and Agents: The Structure of Business 37,

37 (John W. Pratt & Richard J. Zeckhauser eds. 1985).

12

Reliance on an agency relationship for innovation in this way can realize several types of benefits

for a principal, including shifting risks of project failure, adjusting reward patterns to the preferences of

innovation providers, and allocating innovation efforts to the low-risk provider. The features of these

benefits are described more fully in the remainder of this subsection.

1. Shifting Risks of Failure

An agency contract is a means of shifting specific risks of innovation failure from the principal

involved to the agent. If an agent pursues innovation, but is unsuccessful in these efforts and does not

achieve the results that trigger the right to a reward, the expense and opportunity costs associated with the

innovation attempt will be borne by the agent and the principal will be out nothing. In contrast, if the

principal undertook similar innovative efforts on her own, she would bear the risks of failure. Hence,

agency contracts are means to shift risks of innovation failure to agents. Rewards provided to agents must

be sufficiently large to compensate the agents for the risks they take in seeking innovation.26

In this risk shifting, various allocations of innovation risks between agents and principals are

possible by adjusting whether agents are paid for their efforts regardless of their ultimate success in

producing useful innovations. This type of arrangement limiting (or even eliminating) the risks born by

innovation agents is present in many large companies that pursue innovation. Corporate principals

sometimes pursue innovation inside their companies, acquiring all rights to innovations that produced by

their employee-agents. They typically provide those agents with relatively guaranteed salaries and

supporting resources to carry out innovation efforts. If a given innovation project fails under these

circumstances, it is the corporate principal that bears the loss. By contrast, where outsiders attempt to

develop a new innovation and to later convince a corporation to acquire rights in the innovation (usually

to utilize the corporation’s resources to enable large scale production and sales of the innovation) the

outsiders bear the full risks of failures of their innovation efforts.

2. Adjusting Contingent Innovation Rewards to Reward Preferences

Agency contracts can also be means to structure the timing and nature of rewards to

accommodate the specific risk preferences of innovators. This type of adjustment can be achieved by

including within innovation rewards packages various combinations of fixed rewards for innovation

efforts (in which instance innovators only bear the risks of not completing the indicated efforts) and

contingent rewards dependent on the success of an innovation (in which instance innovators share the full

risks of innovation failure).

The objective of optimal reward structuring is to provide rewards to agents carrying out

innovative efforts that are sufficient to encourage cost effective innovation efforts. There are two ways to

accomplish this: 1) linking the payment of rewards to the innovators to the completion of innovative

designs with significant practical value to users of the innovations or 2) linking rewards to efforts by the

innovators that are highly likely to produce useful designs. A strategy that rewards only useful results

delays the "settling" of the innovator's reward to the completion of an innovation project when the nature

of the results and their value to users can be measured with some accuracy. A strategy that rewards

efforts tending to produce successful innovations focuses on rewarding progress, with the advantage of

giving an innovator some interim compensation as a project progresses and as the innovator commits

more and more time to the project. The disadvantage of a strategy rewarding efforts per se is that the

relationship between efforts and results may be poorly understood. As a consequence, efforts may be

26

See generally Steven Shavell, Risk Sharing and Incentives in the Principal and Agent Relationship, 10 Bell

J. of Econ. 55 (1979).

13

rewarded that are ultimately unproductive and that have little impact in promoting a successful innovation

result. Hence, in innovation settings, triggers for agent payments that turn on results achieved for users

rather than mere efforts supporting innovation are typically more likely to drive innovators to consider the

interests of innovation users at all stages of innovation projects. A reward scheme that promises an agent

a meaningful fraction of the net new utility his or her advance will realize for users will tend to cause a

risk neutral agent to pursue efficient types of innovation on behalf of the principal. The agent will tend to

see the interests of the user (in the net new utility) as the agent’s own interest, with the result that the

agent will focus and calibrate his or her efforts in accordance with the utility implications for the user.

Where an agent pursuing innovation is risk averse, having such a party bear the complete risk of

innovation failures will generally cause him or her to pursue a less than optimal level of innovation as a

means to reduce the associated risks. To encourage greater effort without increasing the net risk to a risk

averse agent, a principal can agree to give the agent both a flat fee to compensate for some of the agent’s

risk taking and a further contingent fee to be paid upon production of a successful innovation. The flat

fee component of this arrangement will reduce the innovator’s net risk, thereby matching the risk adverse

preferences of the innovator. Assuming that effort undertaken by an agent can be measured by a principal

in a costless yet accurate fashion, optimal results will follow if 1) the agent is paid a flat sum if the agent

adopts a pattern of effort which is likely to produce a successful innovation and 2) a further contingent

reward is paid if the agent produces a successful innovation with widespread practical implications and

usage corresponding to significant aggregate user gain.27

3. Limiting Rewards to Confirmed Project Success

Another key variable in contractually-defined agency relationships promoting innovation is the

triggering criteria for results-oriented rewards given to an agent and the corresponding strength of the

bond between the interests of the agent and those of the principal. Where an agent’s reward is fully

dependent on delivering an innovation with tangible benefits to the principal and the amount of the

reward approximates the value of the benefits, the interests of the principal and agent will be fully

aligned. The agent will tend to conform his or her actions to those tasks that the agent expects will be

most likely to increase the benefit for principal and enhance the associated reward for the agent.

However, where (due to either poorly defined criteria for granting rewards or poor monitoring of

whether desirable results have been achieved) rewards are paid to agents without desirable results having

been achieved for their principals, there is a substantial chance that potential innovators will take the

minimum actions necessary to qualify for the rewards, but few steps beyond this. Under these

circumstances, less than optimal levels and types of innovative efforts may result as agents pursue the

conduct that will further their own interests, but not necessarily those of their principals. Hence, reward

criteria and monitoring accuracy are always in tension with the success of agency processes for

innovation. As will be seen, patent rights and the commercial market testing for products such rights

enable provide some solutions for these problems.

27

The otherwise limiting effect of an agent's risk aversion is overcome in this type of arrangement by

relieving the agent of an element of project failure risk. Payments rewarding the agent for efforts rather

than results shield these payments from project failure risks and lessen the overall risks of the project to the

agent. Where agents capable of innovation are risk averse – as will often be the case for individuals who

may not want to put all of their income potential and associated wellbeing on the line in connection with a

single innovation project – incentives capable of attracting agents to innovation projects and promoting the

most efficient innovation efforts possible may require the use of a reward schedule that isolates innovators

from some risks of project failure and rewards innovators for completion of success-promoting efforts in

their own right. See Steven Shavell, Risk Sharing and Incentives in the Principal and Agent Relationship,

10 Bell J. of Econ. 55, 56 (1979).

14

4. Allocating Innovation Attempts to Low-Risk Providers

Agency arrangements for innovation are also means for allocating innovation projects to the low-

risk providers of innovation services. These arrangements allow parties with superior expertise in given

fields to apply that expertise in ways that heighten the chances for innovation success over the chances for

success otherwise possessed by a principal. A person with expertise in a field may possess knowledge

that allows him or her to narrow attempted innovation efforts to those with a high likelihood of success.

Alternatively, a person with special expertise or skills may be able to complete innovation attempts more

quickly in a given period, thereby at least testing alternatives quickly and eliminating failures rapidly.

Under conditions where innovation must be completed in a specific time or with limited resources, this

type of increased speed of innovation attempts will tend to increase the chances for success in a particular

innovation project. Either of these consequences of enhanced skill will reduce the expert's risks in taking

on an innovation attempt relative to the risks that would be borne by less expert principals if they were to

attempt a similar innovation project. Indeed, the need for the exceptional skill or knowledge of an agent

in a particular area to direct and complete complex research is often the basis for an agency relationship

aimed at innovation and the willingness of a principal to pay an agent a considerable sum for the latter's

services.

This type of trading on the expertise of an agent with an eye towards the projected value of

products of that expertise is not limited to technical innovation. It forms the bedrock for many

professional fields, including the practice of law. For example, in making legal arguments or crafting

legal documents, a client typically relies on the superior expertise and skill of an attorney to produce

successful results that have specific projected value to the client. While both a client and his attorney

may fail in such tasks, the risks of failure of an inexperienced client in making his own legal arguments or

drafting his own legally significant documents will generally be much greater than the possibility of

failure (that is, a poor legal result) on the part of an experienced lawyer.

Hence, while there is still some residual risk of failure when a complex innovation task is

performed by a party with superior knowledge concerning the task, the shifting of a task from principal to

agent can be a means to allocate tasks to the low-risk provider and thereby reducing the risk of innovation

failure. By reducing the number of failed innovation attempts and resulting losses of resources, such task

shifting and risk reduction (achieved by matching innovation attempts with pre-existing expertise and

skills aiding in those attempts) should produce corresponding increases in the efficiency of innovative

processes.

D. Limiting Breakdowns in Innovation Contracting

A number of factors may impede the formation of incentive contracts establishing agency

relationships for innovation. While all of the relevant reasons for contracting breakdown are not

discussed here, the circumstances addressed in this section are commonly encountered reasons why

contract-based innovation relationships may be difficult to form (and why patent-based substitutes for

these relationships may be particularly valuable).

1. Difficulty in Identifying and Targeting Capable Innovators

Even in the forming the simplest of agency relationships involving one innovator working on

behalf of one potential innovation user, gaps in information about the capabilities of potential innovators

may preclude finding a qualified innovator and lead to breakdowns in contracting processes. Transaction

costs of finding a qualified expert and including that expert in an innovation project may be too high. The

relevant innovation users may lack the information (or the resources needed to generate the information)

15

required to identify the relevant innovators. A disabling information gap may result from either 1) a lack

of information about who has the relevant skills and knowledge to produce a particular type of advance or

2) from doubts on the part of potential innovation users about the veracity of information revealed by

innovators about their qualifications (as where experts assert a level of expertise but lack a track record of

proven results that would confirm a likelihood of being able to produce a desired type of advance).

Without sufficient information signaling who has the needed abilities and, therefore, who to contract with,

potential users of advances will tend not to form agency contracts for completion of innovation projects.

A system of patent-based innovation efforts will tend to overcome this problem by simply

creating contingent rewards for successful innovation, with the size of the rewards scaled to the value of

innovations to users, but with no need for users to contact innovators in advance of innovation projects.

Potential innovators will see the relevant contingent rewards and self-identify themselves by starting

innovation efforts under this system. Innovators will tend to reveal (and, indeed, advocate) the value of

their invention results via commercialization efforts that portray the uses and value of advances.

Innovators will be motivated to both increase and reveal the value of their advances because patent-based

rewards are maximized by creating commercially valuable advances that are popular with users. Initial

information uncertainty is resolved by performance. In this regard, the patent system establishes an

equivalent of a unilateral contract offer that can be accepted by any party producing a commercially

valuable advance that has the features such as novelty and non-obviousness required to qualify for a

patent. The patent process is the economic equivalent of a bounty system, with the bounty payment

realized through the combination of creation of a new advance and commercialization and popularization

of that advance to place it into the hands of many paying users.

2. Overcoming Problems in Forming Group Contracts

Even where a relevant innovator with sufficient technical skills is identified, formation of a

contract incentivizing this party to pursue a useful advance on behalf of multiple potential users may be

difficult because of strategic behavior among the users. Each user may hold back, hoping to get other

users to pay for innovation expenses and then using the resulting advance at a low cost. Ideally, where

multiple users share a need for a particular type of advance, they should share the cost of subsidizing the

research needed to satisfy this shared need. The group will benefit if the aggregate costs of research do

not exceed the aggregate incremental value of an advance to the group members. A payment scheme that

entailed joint contributions to such a research project up to the level of their perceived potential gains

would allow a group to act as a single principal in a contractually-defined agency relationship with the

targeted innovator.

Unfortunately, the formation and administration of this type of group contracting for innovation

will often be problematic. Both information gaps and strategic behavior may prevent potential invention

users with like invention needs from finding each other and presenting joint incentives to a contractually

engaged innovator serving the user group.

a. Finding Relevant Group Members

Transaction costs in identifying innovation users with similar technological needs will hinder

such group contracting by limiting the scope of invention user groups to less than the full range of

potential users. Some interested users will not be found and will be omitted from potential contracting

groups, leading to amounts of contractual incentives for innovation that are smaller than if all potential

users’ interests were reflected. By excluding some users, partial user groups will also exclude some

incentives for innovation. The resulting incentives will be less than the optimal rewards as measured by

the full potential usefulness (and societal value) of an advance under development. With the less than

optimal incentives imbedded in innovation contracts (or contract offers), some desirable innovation

16

projects will not be initiated (because they seem not to have net value given the limited number of parties

organized to pay for them) or will be pursued with less than optimal levels of resources and efforts

(because they are cost-effective to complete in a limited form, but not with the full level of efforts that

funding by all of the interested users would have supported).

To illustrate this point, consider an advance that has projected usefulness to approximately 10000

similar businesses. For simplicity, assume that each of these businesses would expect to gain

approximately $1000 from the advance. If all were involved in a contracting effort to encourage creation

of the advance, the users as a group would be willing to collectively pay a bit less than $10,000,000

($1000 x 10000) to reward someone for development and delivery of the advance. A contract

establishing a promised payment of about $10,000,000 upon successful delivery of the advance would be

cost-justified at the group level (in the sense that the group would receive more back in aggregate value

gained than it paid out in development costs for the innovation effort).

However, if contracting problems preclude banding together all of the interested firms in this

way, a fraction of the group contracting for the same innovation might offer only a smaller payoff upon

successful production of the advance. This smaller promised reward would tend to prompt less

substantial efforts to produce the advance. For example, suppose that (even though there are actually

10,000 similarly situated firms) a party needing an advance only identifies and reaches a joint-

development agreement with 9 other firms that also need the advance. Managers of the first company

work in conjunction with the 9 other firms to establish a joint contracting effort seeking design and

production of the advance in question. The maximum reward that will be cost justified for the 10 firms to

offer is $10,000 ($1000 x 10). The incentives created by embedding terms for this type of contingent

contractual reward in a contract for production of the advance will encourage only a fraction of the

innovation development effort that a more inclusive contract and larger reward would incentivize. Better

information leading to the involvement of more potential invention users in the payment of innovation

rewards will tend to produce larger contingent rewards and incentivize greater innovation efforts. By

contrast, information constraints limiting the identification of potentially interested users will impair the

formation of contracting groups and correspondingly restrict the effectiveness of these groups in

promoting innovation efforts by agents at efficient levels reflecting the full potential value of advances to

their complete set of users.

This type of information limitation on identifying contracting parties will tend to impair the

formation of complete innovation contracts and the creation of optimal contractual incentives for

invention where the full set of potential users for an advance is either 1) numerous, motivated by diverse

product detail preferences, or isolated in separate industries or operating environments. In these

situations, identifying parties with like needs for a specific advance may be particularly difficult as the

necessary fact finding will extend across numerous situations or parties. Where the resulting information-

gathering barriers are large, the formation of complete contracting groups for the incentivizing of

innovation will be correspondingly hard and frequently unsuccessful. In these settings where contractual

incentives are likely to be absent or at low levels not reflecting the full value of potential advances, the

need for patent incentives as substitute sources of innovation incentives will be particularly great.

b. Group Formation Problems Due to Negotiation Difficulties

Even if the similarly situated parties with joint interests in a new advance can be identified, some

of these parties may refuse to participate in a group innovation contract (or seek to participate without

paying their share of innovation incentive costs) due to problems in identifying and sorting out the

relative roles (and payment obligations) of the group members in a joint innovation project. These relate

to problems that will surface as the potential group members seek to negotiate their specific roles and

contractual obligations in carrying forward an innovation project.

17

Negotiations over a group project may founder over the specification of many types of inter-party

terms, including disagreements between the parties about how much benefit each will obtain from an

innovation being sought and how much each should pay for its development. Each party may attempt to

downplay the gain that they will receive from the innovation at issue in hopes that their fellow

contributors would bear most of the reward "costs" of producing the innovation. These sorts of strategic

choices of the potential contracting parties may keep them from reaching agreement on appropriate

reward levels. As with information barriers that precluded the formation of innovation contracts with

incentives reflecting the full value of an advance to potential users, strategic behavior that similarly limits

the involvement of all innovation users and the amounts that the contracting users will pay to a successful

innovator are likely to impair or preclude the development of some advances with net social value (taking

into account the aggregate benefits to all potential users).

c. Group Formation Problems Due to Strategic Behavior

The formation of a complete contracting group promoting innovation by an agent may also be

impeded due to the strategic behavior of potential group members. Some of the parties potentially

benefitted by an advance may refuse to participate in a group innovation contract (or seek to participate

without paying their share of innovation incentive costs) due to strategic behavior aimed at gaining an

advantage over their fellow group members. Each interested party may follow a strategy of trying to get

as much return from an advance as possible by seeking to throw the main costs of innovation rewards on

the other contracting parties in the group. Each potential user will have incentives to seek access to

innovations without sharing in the innovation costs, meaning that there will be strong advantages to

holding back, not joining the innovation contracting groups (or nominally joining these groups but

avoiding performance of obligations to pay portions of innovation rewards). If successful in these efforts,

the users will obtain the benefits of innovation as “free riders” who have not paid for the associated costs

of innovation. This type of strategic maneuvering may preclude the formation of a group innovation

contract or lead to one that involves so few potential users as to severely under fund and under incentivize

the resulting innovation efforts.

5. Costs of Defining and Administering Group Innovation Contracts

Assuming that a group of parties with shared interests in potential type of innovation are willing

to attempt a group innovation contract with an agent, several types of contracting difficulties may arise

and consume resources that would otherwise be devoted to the creation of incentive payments for

innovation. These problems may also reduce innovation rewards below optimal levels (taking into

account the full range of benefits available from a particular innovation). The following types of

problems may arise in defining and administering the terms of innovation contracts involving groups of

potential innovation users who are obligated to pay innovation rewards to agents producing advances.

a. Disputes Over Contract Terms

The negotiation of specific terms of a group innovation contract may be difficult because this

process will force group members to agree on terms that will often be in dispute. Negotiation of essential

contract terms may force potential innovation users to focus on differences in the characteristics and

interests of the various users and their corresponding desires for different contract terms. Differences of

interests – and corresponding differences in desired contract terms for innovation contracts – may concern

such key topics as the nature of the innovation being sought, the circumstances constituting successful

contract performance under which contingent payments will be due as innovation rewards, the

administrative arrangements under which contract participants will pay portions of innovation rewards,

and the types of information about and control over an innovation that contracting parties will have if a

18

successful innovation is produced. Even if parties with diverse characteristics can come to agreement on

these key features of a group innovation contract, the specification of such complex terms among a

substantial number of potentially diverse users will require difficult and costly transactional steps, with

corresponding decreases in the amounts of innovation rewards and reductions in the effectiveness of

resulting innovation incentives.

b. Contract Administration Costs

Assuming that a group contract can be established calling upon an innovator to produce an

advance on behalf of a group of potential users, the provisions of such a contract must still be

administered. Both measurement and administration costs associated with the incentive payment

processes established under an innovation contract may also undercut some of the value of such contracts.

Amounts spent on these tasks will reduce the amounts devoted to innovation incentives with a resulting

lessening of incentives below efficient levels.

Several steps in the administration of group contracts for innovation will require the completion

of difficult and costly conduct and item measurement. Parties contributing portions of contingent

payments under innovation contracts will want their contributions and those of other parties carefully

accounted for in accordance with their joint contract terms. The willingness of each to make their own

payments will depend in part on the firmness of the accounting mechanisms and the trust that they place

in those mechanisms. Hence, these measurement and administrative processes will need to be substantial

and will be likely to contribute significant costs to the administration of innovation contracts involving

multiple innovation users and incentive payers.

Similarly, significant measurement steps and costs may arise in carrying out the terms of

contingent innovation contracts. The success of an innovator working under such a contract must be

monitored carefully to determine if he or she is entitled to the contingent innovation rewards specified in

the contract. This will sometimes involve expert analyses of innovation products, with corresponding

fees paid to the experts. Both the innovator and the potential users who are parties to a group innovation

agreement will have strong interests in having evaluations of innovation results performed thoroughly and

fairly. Potential innovators will fear that poorly performed evaluations will improperly deprive them of

innovation reward payments (after having incurred considerable efforts and expense to produce a

successful advance). Potential users will worry that inaccurate evaluations of innovation results tendered

under a contract will result in reward payments for what were really inadequate submissions.

Erroneous evaluations that result in over- or under-payments of invention rewards will undercut

the effectiveness of contract-based agency processes in aligning innovators’ interests with those of

innovation users. This type of misdirection of agency efforts – producing complacency in innovators who

can expect payments without actually producing successful innovations and deterring diligent efforts by

innovators who fear non-payments despite successful performance – will tend to undercut the success of

contract-based agency processes for innovation. Even where insecurity over contract payments is

overcome by careful evaluations of innovations as preliminaries to payment commitments, these careful

evaluations will add costs to the administration of innovation contracts and reduce potential innovation

rewards accordingly.

E. Patent Rights as Substitutes for Agency Contracts in Cases of Contracting Breakdown

Against the backdrop of contractual agency arrangements just described (and the difficulties in

forming and administering these arrangements), the potential advantages of patents as vehicles for non-

contractual agency relationships furthering innovation becomes clear. Patents can be incentive creators

when contracts cannot be. Patent-influenced processes promoting innovation create agency relationships

19

in which legally-created promises of patent rights and rewards link the interests of invention providers to

invention users. Patent rights establish incentives that make innovators agents of principals who are the

users of innovations. Patent rights create the incentives for innovators to target, perfect, commercialize,

and distribute inventions that are useful and popular with numerous users. Innovators are encouraged to

think like agents of invention users, seeking always to answer the question “How can I produce an

advance that will provide the greatest beneficial results for users?” The notion of what constitutes a

beneficial result – including such features as item cost, ease of use, and value of results obtained – will be

measured from the perspective of users, not innovation producers. Hence, patent rights (and the

commercial value in advances that they create) cause innovators to perceive invention users’ interests as

the innovators’ interests as well. Patent rights align these interests in the same ways as contract incentives

align the rights of innovators and invention users under innovation-focused agency contracts.

By using what we already know about the dynamics of agency contracts (from the extensive

study of contractual agency relationships to date), we can project many agency features of patent-

mediated agency processes. By modeling patent incentives and their agency consequences in terms of

their contract-based counterparts, we can gain valuable insights into the ways that patents influence

innovation. We can also use this analytic perspective to better tailor the scope and force of patent rights

to optimize agency efforts leading to valuable innovation.

Patent rights are most needed as substitutes for agency contracts promoting innovation in

situations where the formation of agency contracts for innovation is problematic and unlikely.28

While

contracts are capable of more careful and detailed tailoring to the circumstances and needs of particular

innovation users, where contracts incentivizing innovation are unlikely to be formed (or will be formed

with only a fraction of the relevant beneficiaries, potentially leading to far less than optimal incentive

payments), other innovation incentive mechanisms may play a valuable role.

Patents provide a means to extend important innovation incentives and agency relationships

beyond the realm of contracts. The incentives created by patents may be best thought of as contractual

substitutes for use as innovation drivers in settings beyond the realm of likely formation of agency

contracts for innovation. Put simply, patent rights are solutions for innovation promotion in cases of

contracting breakdown between innovators and innovation users. Ideally, to fulfill their potential as

contractual substitutes, patent rights will establish contingent payment terms for successful innovators

that approximate the contractual payment terms and incentives that innovators and innovation users

would choose if they could enter into contracts will perfect information and no transaction costs or

strategic behavior. Under this model, patent rights should be triggered (and have projected scope and

commercial value) similar to the incentive and performance terms that principals who are potential

innovation users would offer to potential innovators where the principals able to contract the innovators

and offer the terms.

III. Expanding Innovation Analyses via Agency Theory

28

Patent rights may serve a valuable function even where some contractual incentives for innovation are

available. Assuming that both contractual and patent incentives are available for the production of a given

type of invention, these incentives may not be equivalent. Contract incentives created by a particular user

or group of users will reflect the needs and invention value of an advance to just these users. The

incentives created (and the force on innovators and innovation efforts) will be curtailed and biased

accordingly. By contrast, the incentives promised by patent rights will tend to reflect the full range of uses

and commercial value of an advance since these can be controlled once a patent is in place. In this respect,

patent rights create broader and more complete incentives than most contractually created innovation

incentives even where both are available.

20

A. Agency Theory as an Analytic Tool

Agency theory provides an organized means to summarize and study the salient features of

situations (including patent-mediated innovation) where one or more parties (the agent or agents) act to

further the interests of one or more other parties (the principal or principals).29

Agency theory provides

useful accounts of the causes and consequences of goal incongruence in principal and agent

relationships.30

It can also assess how information flows and risk allocations affect the effectiveness of

agency relationships.31

In applying agency theory, it is important to remember that this theory is not limited to

contractually created agency relationships. An agency relationship capable of assessment through agency

theory is present wherever one party controls an outcome of interest to a second and the first party's

efforts are incentivized via contingent rewards tending to encourage the first party to pursue the interests

of the second.32

Hence, non-contractual, but patent-influenced relationships in which innovators are

incentivized to advance the interests of invention users are instances of agency relationships that can be

evaluated and improved using agency theory.

Much of the agency theory literature has focused on three aspects of principal-agent relationships:

(1) the preferences of principals and agents in these relationships (including means to alter the preferences

through contingent reward provisions), (2) the uncertainties present in principal-agent relationships and

their impacts on participants in these relationships, and (3) the information flows within principal-agent

relationships and from surrounding environments that influence the success of these relationships.33

Studies have also examined means for risk sharing within principal-agent relationships, the features of

29

For overviews of agency theory and the types of problems the theory can address, see Hal R. Varian,

Intermediate Microeconomics: A Modern Approach 667-88 (6th ed. 2003); Kenneth J. Arrow, The

Economics of Agency, in Principals and Agents: The Structure of Business 37-51 (John W. Pratt & Richard

J. Zeckhauser eds., 1985); Kathleen M. Eisenhardt, Agency Theory: An Assessment and Review, 14 Acad.

Mgmt. Rev. 57 (1989). A number of scholars have used agency theory to analyze legal problems. See,

e.g., Howell E. Jackson et al., Analytical Methods for Lawyers 88-100 (2003); Robert H. Sitkoff, An

Agency Costs Theory of Trust Law, 89 Cornell L. Rev. 621 (2004); Eric A. Posner, Agency Models in Law

and Economics, in Chicago Lectures in Law and Economics 225 (Eric A. Posner ed., 2000); Easterbrook,

Insider Trading as an Agency Problem, in Principals and Agents: The Structure of Business 81, 87-98

(John W. Pratt & Richard J. Zeckhauser eds. 1985).

30 See J. B. Barney & W. S. Hesterly, Organizational Economics: Understanding the Relationship Between

Organizations and Economic Analysis, in Handbook of Organizational Studies 115, 115-147 (S. R. Clegg,

C. Hardy, & W. R. Nord eds. 1996).

31 See K. M. Eisenhardt, Agency Theory: An Assessment and Review, 14 Academy of Mgmt. Rev. 57, 57-74

(1989).

32 See Steven Shavell, Risk Sharing and Incentives in the Principal and Agent Relationship, 10 Bell J. of

Econ. 55, 55 (1979)(noting that contractually-induced agency relationships are only a subset of all such

relationships and that the essential features of agency relationships (possessed by both contractual and non-

contractual agency relationships) are that in an agency relationship 1) a principal enjoys the outcome of an

agent's actions, 2) the success of the agent's actions in promoting the principal's interests is determined by a

combination of the agent's efforts and additional uncontrolled circumstances not controlled by the principal

or agent, and 3) to encourage favorable efforts, the principal commits to pay the agent a fee upon the

completion of efforts or the attainment of results that advance the principal's interests).

33 See Michael C. Jenson, Organizational Theory and Methodology, Accounting Rev., April 1983, at pp. 319-

339.

21

optimal contracts between principals and agents, and welfare comparisons of equilibrium contracting

solutions to agency relationship formation with and without information costs.34

Additional studies have used agency theory to predict the impacts on principal-agent relationships

of changes in surrounding environments. Changes considered in these studies have included alterations in

contracting environments, changes in the technologies used to monitor agents’ actions and trigger

performance rewards, and shifts in the bonding arrangements used to signal commitments by agents to

particular courses of conduct.35

Similar studies have also examined the impacts on principal-agent

relationships of variations in the capital intensity of surrounding business enterprises, the specialization of

business assets, information costs, features of capital markets, and characteristics of labor markets.36

An additional theory related to agency theory has also been applied to evaluate corporations and

other organizations that operate through the actions of multiple agents. This theory treats an organization

or "firm" as the nexus of numerous agency relationships. The formation of principal-agent relationships

within a given firm (and the specification of the terms of those relationships) are treated as the results of

numerous agency-tailoring decisions. Ideally, the resulting set of principal-agent relationships will

constitute an equilibrium condition in which the firm’s agency features maximize the operating benefit to

the corporate organization. Analyses employing this theory focus on both the means to make particular

principal-agent relationships more effective in corporate business activities and further means to adjust

the terms and content of one corporate agency relationship relative to another.

Finally, some agency studies have evaluated agency relationships constructed and administered

through non-contractual processes. In these processes, a party or body (such as the government) creates

agency relationships by setting up contingent rewards through non-contractual means. The contingent

rewards are constructed so as to encouraging one party to act on behalf of another. Parties tend to

respond to promises of non-contractual rewards much like they respond to similar contract-based rewards.

Where the payment of the rewards is tied to completion of actions of interest to particular principals,

parties in both circumstances tend to act as agents and align their conduct goals with the interests of the

principals. Hence, many features of agency theory developed for evaluations of contract-based

relationships can be used to study and evaluate agency relationships created via non-contractual means.

Patent-mediated innovation relationships linking innovators and innovation users are examples of

non-contractual principal-agent relationships. Patent rights create the contingent rewards that energize

agency behavior. Unfortunately, non-contractual agency arrangements such as these have rarely been

assessed as agency processes. Hence, it will frequently be helpful to look to studies of contract-based

principal-agent relationships for insights about agency dynamics and design features that can be carried

over into the less well understood area of patent-influenced principal-agent relationships.

B. Key Characteristics of Principal-Agent Relationships

1. Establishing Agency Incentives and Aligning Principal and Agent Interests

For purposes of agency theory, each party acting as a principal or agent in a principal-agent

relationship is modeled as a rational actor pursuing his or her own interests. These interests are defined

by the values of the participants in the relationship, background environmental features surrounding the

34

See id.

35 Id.

36 Id.

22

relationship, and the rewards and incentives created by the terms of the relationship.37

Each participant's

perception of his or her self-interest and how to pursue it is also shaped by the information set available to

that individual.38

Acting within an agency relationship, an agent controls the nature of his or her conduct and,

hence, the degree to which the conduct furthers the interests of the principal. These conduct choices are

fundamental features of the relationship. A principal can influence these choices – and how closely the

agent’s conduct adheres to the interests of the principal – by adjusting the amounts and triggers for

rewards or fees that encourage an agent to undertake particular efforts within an agency relationship.39

Payment of rewards or fees can be triggered by attainment of specific results desired by principals (such

as occurs with sales commissions paid to company employees for completion of specified volumes of

product sales) or for the completion of efforts that are steps towards results desired by principals (such as

occurs with salaries paid to personnel in the advertising department of a company whose work on

advertising efforts is hoped to support sales transactions elsewhere in the same company).40

If the specific results desired by a principal are known at the outset of an agency process (or at

least the actions by an agent that the principal projects will be most likely to serve the principal's goals are

known), several additional features of a principal-agent relationship can be adjusted and refined to best

assure attention by the agent to the preferred result or action. Adjustable relationship features include: 1)

the minimum rewards needed to get an agent's threshold attention and to cause the agent to consider how

to pursue the principal's activities, 2) the amount of contingent rewards payable upon attainment of

desirable results or upon the completion of efforts tending to produce the results, 3) the circumstances

triggering payment of rewards to the agent, 4) means for gathering information on the agent’s

performance and for evaluating that information and transmitting the evaluations to the principal, 5)

mechanisms for gathering information on circumstances surrounding the agent's actions (to the extent that

this circumstantial information is relevant in evaluating the appropriateness of the agent's actions), 6) the

risk aversion levels of the principal and agent, 7) the allocation between principal and agent of risks of

failure of agency-related activities, and 8) additional payments to agents to encourage them to bear risks

37

See K. M. Eisenhardt, Agency Theory: An Assessment and Review, 14 Academy of Mgmt. Rev. 57, 57-74

(1989)(noting that, in agency theory analyses, principals and agents are typically assumed to be 1) subject

to bounded rationality, 2) self-interested, 3) opportunistic, and 4) risk averse).

38 See id.; Stanley Baiman, Agency Research in Managerial Accounting: A Second Look, 15 Accounting Org.

& Soc. 341, 342 (1990).

39 In general, principals can align agents' interests with those of the principals and minimize agency costs by

offering the agents a share of the benefits achieved through the agents' actions. See J. B. Barney & W. S.

Hesterly, Organizational Economics: Understanding the Relationship Between Organizations and

Economic Analysis, in Handbook of Organizational Studies 115, 115-147 (S. R. Clegg, C. Hardy, & W. R.

Nord eds. 1996).

40 Outcome based arrangements -- in which the fee paid to an agent depends on the attainment of an outcome

that is desirable from a principal's perspective -- tend to align the interests of principal and agent.

However, these types of fee schemes sometimes make agents more risk averse by putting them at risk of

failures due to factors outside their control – that is, by putting their attainment of fees within the agency

relationship at risk of failures by other parties contributing to the outcomes that will trigger the fees. By

contrast, fee schemes that provide for payments to agents based on efforts tending to promote successful

results regardless of whether those results are attained tend to promote greater risk taking by agents since

they need not be concerned over risks outside of their own actions. See K. M. Eisenhardt, Agency Theory:

An Assessment and Review, 14 Academy of Mgmt. Rev. 57, 57-74 (1989).

23

of failure regarding the pursuit of the principal's interests that the agents would not otherwise be willing to

accept.

Agency theory models the operation of agency relationships in terms of these (and other) agency

features. Agency theory helps to define the incentive problems in these settings which will be relevant in

considering the optimal innovation incentives potentially derived from patent laws. In addition, agency

theory provides means to predict the impacts on agency behavior of various adjustments to agency terms,

including the types of changes in agency relationships that may follow from alternations in patent law

standards and resulting innovation rewards.

To see how agency theory can help us understand patent-mediated agency processes, it is useful

to consider the key features of a simple agency relationship as seen through the lens of agency theory.

Agency theory helps to reveal the important functional features of such a relationship, suggesting the

levers and problems governing the operation of such a relationship. Consider the following simple

agency arrangement.41

A real estate agent is engaged to sell a home by the home owner under a contract

calling for the agent to receive a commission of 5 percent of the sales price upon successful completion of

the home sale. Under these circumstances, the agent is encouraged to undertake a specific range of cost-

effective actions, with the measure of cost-effectiveness measured from the agent’s perspective. For

example, the agent will have an incentive to take an action that costs $1 and that improves the projected

price realized for the seller by $100 since the agent will predict an increased payoff (and a perceived

increase in the agent’s own welfare) of $5 from this action. However, the agent will not have an incentive

to take an action that costs $10 and increases the seller's price by $100 since the agent will only gain $5

from this action.

The lack of motivation of the agent to take this last action reflects some degree of interest

divergence between the principal and agent – that is, a gap in ideal incentives which, if fully successful,

would completely align the interests of principal and agent. The action costing $10 would have advanced

the seller's interest since the seller's gain was greater than the cost, yet the agent has no interest in

pursuing this action. Such a lack of interest is evidence that the fee arrangement between the seller and

agent has not perfectly aligned their interests, producing a welfare loss in the operation of the agency

arrangement. This welfare loss corresponds to the failure to take the $10 action that would have resulted

in a sale of the house for $100 more and a net increase in welfare for the home owner of $100 (which

could have been shared with the agent under a different fee structure to compensate the agent for his

actions and still have realized a net welfare gain of $100 - $10 = $90). The losses that stem from such

misalignments of interests are called agency costs.

2. Agency Costs Impairing Agency Processes

Three types of agency costs generally plague agency relationships and processes. The different

types of costs stem from flaws in different components of agency relationships. The three types of

agency costs stem from principals’ difficulties in monitoring agents’ actions (monitoring costs), agents’

difficulties in convincing principals of the agents’ capacities and trustworthiness (bonding costs), and

agents’ inattention to some aspects of principals’ interests resulting in reductions in principals’ welfare

below ideal levels (residual costs).

41

This example is described in Frank H. Easterbrook & Daniel R. Fischel, The Economic Structure of

Corporate Law 91 (1991). See also Frank H. Easterbrook & Daniel R. Fischel, Corporate Control

Transactions, 91 Yale L.J. 698, 701 (1982); Eric A. Posner, Agency Models in Law and Economics, in

Chicago Lectures in Law and Economics 225, 225-29 (Eric A. Posner ed., 2000).

24

Monitoring costs and bonding costs are essentially administrative costs of operating agency

relationships born by principals and agents respectively. Even in relationships where perfect information

was provided to principals and agents, the costs of providing this information would be a drag on

operations and reduce the efficiency of these relationships. For principals, costs spent in monitoring the

activities of agents and determining when reward payments to these parties are due will reduce the net

value of operating through agency arrangements and correspondingly reduce the amounts that principals

are willing to pay to agents as reward incentives. For agents, the amounts that they must expend to

establish their reputations and retain the confidence of principals are typically expenses that they must

bear, meaning that the costs of these tasks will reduce the net value of agency rewards they perceive and

receive for being part of agency processes. Both these type of administrative expenses will have the net

effect of reducing the net amount and functional impacts of agency rewards, with the ultimate result that

some agency relationships that would have been formed and effective without these costs will not be

established in light of the reduced value resulting from such costs.

Residual costs are essentially opportunity costs corresponding to actions furthering principals’

interests that were encouraged by the net incentives presented to agents and within the capacity of agents,

but nonetheless not taken by the agents. Residual costs typically result from a misalignment of the

interests of principals and agents. The $100 loss resulting from the real estate agent's failure to pursue

actions in the house seller's interest in the scenario discussed above is an example of a residual loss due to

the lack of full alignment of the agent and seller's interests. Understanding residual costs in agency

contexts requires a consideration of the sources of interest divergence in agency relationships.

3. Sources of Divergence of Interests of Agents and Principals

Divergences of interests between agents and principals may stem from two different types of

weaknesses in the ability of principals (or others acting on their behalf) to motivate agents to pursue the

interests of principals. First, a principal may inaccurately monitor an agent's performance, resulting in

payments of rewards to agents for performance that does not actually serve principals’ interests. If agents

can count on this type of imperfect monitoring of their performance, their incentives will extend only to

the types of performance (and performance features) that can be detected by principals and that will

influence the payment and size of rewards for the agents’ performance. This type of problem is a version

of moral hazard42

-- that is, a hazard or risk that the agent will act in a manner diverting from the interests

of the principal because this divergence is not measured and does not affect the rewards or incentives the

agent receives within the agency relationship.43

For example, this type of moral hazard leading to flawed agency incentives is present where an

employee is given a bonus for completing a certain number of sales calls but the quality of the agents’

42

See generally Bengt Holmstrom, Moral Hazard and Observability, 10 Bell J. Econ. 74 (1979).

43 The classic example of moral hazard arises in the context of insurance where coverage against a loss may

influence an insured party to increase risk-taking behavior because, if a loss occurs with an insurance

payoff, that party will not bear the consequences of the loss. Hence, a party insured for fire losses may be

encouraged to be lax with fire safety because, should a fire occur, the resulting insurance payment will

mean that the insured party does not bear the adverse consequences of the fire. Because insurance relieves

him of his incentive for care towards the prevention of fires and the avoidance of harm to others, insurance

can have the unexpected effected of making an insured party a poorer agent of community members

concerned about preventing fire losses. See generally Kenneth J. Arrow, The Economics of Agency in

Principals and Agents: The Structure of Business 37, 37-51 (John W. Pratt & Richard J. Zeckhauser eds.

1985); R. C. Merton, An Analytic Deviation of the Cost of Capital of Deposit Insurance and Loan

Guarantees: An Application of Modern Option Pricing Theory, 1 J. of Banking and Finance 3, 3-11 (1977).

25

actions in making the calls is not monitored. The calls may have been performed in a perfunctory manner

that was not likely to increase product sales for the employee’s company. In this setting, sales calls are

not of interest to the employer per se, but the completion of these calls is used as an easily measured

surrogate measure for promotion of sales success. Since the performance of company employees is

measured in terms of sales efforts (i.e., numbers of sale calls completed), employees are motivated to

undertake the measured actions (completed the sales calls), but not to take the further steps needed to

complete product sales. Indeed, if their compensation turns primarily on the number of sales calls

completed, they will tend to be discouraged from taking these further steps to conclude product sales

since they consume time and effort, but do not increase employee compensation.

Divergence of interests of principals and agents can also result from a second type of problem

involving “adverse selection" of conduct by agents. “Adverse selection” in this context refers to choices

by actions by agents that do not serve principals’ interests or that even hurt those interests. This may

occur inadvertently, for example, where agents are assigned tasks that they are not qualified to carry out

and make poor selections of actions in stumbling through their activities as agents. More serious (and

potentially more carefully concealed) adverse selection of agents’ actions may occur where agents decide

to carry out actions for principals via fraud or other misconduct. This will occur, for example, where

agents seek to superficially qualify for rewards within agency relationships by taking actions (such as

selling products through fraudulent practices) that qualify the agents for rewards (such as sales

commissions), but which ultimately hurt their principals.

An agent may have opportunities to choose actions that poorly serve the interests of a principal

where the principal does not have sufficient information to detect this type of adverse choice. A principal

may not be able to properly interpret the actions of an agent where the actions take place in a context that

is not fully understood by the principal. Where a principal is not aware of the choices of action available

to an agent, then the principal will sometimes be unable to detect that an agent has chosen activities that

are less than optimal in furthering the principal's interests. In essence, this is an information disparity

problem that stems directly from the information collection and understanding gap that often exists

between a principal and agent even under the best of agent monitoring arrangements. Divergence in the

interests of a principal and agent may result from this information gap due to undetected choices of

undesirable actions by an agent.44

In the sales context just discussed, for example, if an employee's compensation increases based

on the number of sales calls he makes, a typical employee will diligently make and report sales calls in

order to boost his earnings. However, an employee may boost his numbers of measured sales calls by

conducting some calls, fabricating records for additional calls, and reporting both the real and fake calls

as completed efforts. The employer involved may not have sufficient information about the employee's

actual sales efforts to appreciate that the employee is sometimes shirking his duties and falsely reporting

sales activities. The employee’s adverse selection of conduct in failing to pursue some sales calls and

falsely reporting completion reflects a divergence of his interests from those of his employer that may go

undetected and, hence, undiscouraged by the contingent rewards provided to the employee under the

employer’s compensation system.

C. Characteristics of Optimal Principal-Agent Relationships

1. Factors Influencing Incentives and Actions

44

For example, adverse selection occurs when individuals sometimes conceal their illnesses as they seek

health insurance. See M. Rothschild & J. E. Stiglitz, Equilibrium in Competitive Insurance Markets, 90 Q.

J. of Econ. 629, 629-50 (1976).

26

In economic terms, the factors influencing the incentives and actions of principals and agents are

summarized by the following equation defining the expected value of actions to a principal (EU(a)) and

an agent (EV(a)):45

EU(f,e) = ∫ ∫

and

EV(f,e) = ∫ ∫

where:

U(a) = principal's utility function (in wealth)

V(b,e) = agent's utility function (in wealth and effort)

e = agent's effort

z = principal's observation of e

x = outcome

f(x,z) = fee paid by the principal to the agent (a function of x (outcome) and z (observation of

effort))

r(x;e) = probability density of x given e

q(z|x;e) = probability density of z given x and e

An optimal fee schedule within this construct will be one in which there is no alternative schedule that

would make one or both of the parties better while not harming the other party. That is, the aim in this

situation is to select a Pareto optimal fee schedule under which the expected utility for the principal and

agent cannot be improved without harming the other. In formal terms, this means that the first-best fee

schedule is one which solves the problem:

max EU(f,e) over f

subject to

EV(f,e) > or = Vmin

EV(f,e) is maximized over e

where Vmin = the fallback utility level that the agent could achieve by undertaking other types of activities.

This utility is measured by the agent's bargaining power and market forces which dictate the value of the

agent's services in other contexts.

2. Terms of Efficient Incentive Fees

a. Risk Neutral Agents

45

See Steven Shavell, Risk Sharing and Incentives in the Principal and Agent Relationship, 10 Bell J. of

Econ. 55, 57-58 (1979).

27

Shavell has shown that, in these circumstances, the choice of an optimal fee schedule linking the

interests of a principal and agent depends on whether the latter is risk neutral or risk averse.46

Where an

agent is risk neutral, the optimal approach is to adopt a fee structure that depends only on the results

achieved by the agent and which gives the agent all of the resulting benefits from a given outcome, minus

a constant amount kept by the principal (the changes needed if an agent is risk averse rather than risk

neutral are considered in the next subsection). Thus, the optimal fee f for a risk-neutral agent is described

by:

f(x,z) = x - k

where x is the value of the outcome x and k is a constant amount kept by or transferred to the principal.

The latter reflects a minimum payment to the principal to induce the latter to administer and participate in

the applicable principal and agent relationship. Under this type of fee structure, the marginal fee of the

agent and incentive perceived by the agent as an encouragement to invest effort that the agent believes

will affect an outcome will vary directly with the increase achieved in the utility of the outcome. To

encourage increased efforts in proportion to their impacts in improving the outcome of the agent's

activities, the aggregate fee or reward motivating the agent should approximate the full benefit of a

particular outcome, less a minimum fee to the principal.47

Variations in the accuracy with which the agent and principal can identify successful results will

not affect the form of the optimal solution so long as errors in this regard are symmetrical and the net

benefits and rewards are maximized under the same circumstances as would produce the maximum

benefits in the case of perfect information.

However, differing abilities of a principal and agent to perceive successful results will affect the

optimal fee schedule as this difference will lead to increased risks to one or other of these parties.

Assuming that a principal can perceive a successful result with perfect accuracy and administers a fee

accordingly, the relative inability of the agent to perceive the successful result adds a new component of

risk to the agent's behavior. Under these circumstances, an agent may feel at risk that some results

potentially qualifying for greater rewards from the principal will not be perceived by the agent as such

and will not be pursued, thereby adding some increased risk of failure and forfeiture of a fee to the agent's

situation. An otherwise risk neutral agent, once aware of this increase source of risk, will become the

equivalent of a somewhat risk averse agent with perfect information. The nature of the fee incentives

necessary to induce optimal conduct on the part of a risk adverse agent are described below.

On the other hand, where an agent (at the time of fee settlement) has a superior ability to perceive

a successful result in comparison with a principal, then the agent may also perceive an increased risk of

non-payment of a fee. In this combination of circumstances, the source of risk is a bit different. The

agent may believe that she will attain a successful result in some task, but that the principal will fail to

recognize that this result has been attained and fail to pay the relevant fee. Here the source of risk is not

the failure to perceive the course of conduct that will best serve the principal, but rather a justified

46

The case where the agent is risk preferring is basically the inverse of the treatment of the risk averse agent.

Just as a risk-averse agent may need a payment for success-promoting efforts to limit the agent's risk

bearing and encourage efficient pursuit of beneficial outcomes, a risk-preferring agent may be willing to

accept a more of the costs of success-promoting actions in exchange for a chance at a larger payoff upon

achievement of a successful result.

47 Steven Shavell, Risk Sharing and Incentives in the Principal and Agent Relationship, 10 Bell J. of Econ.

55, 64-65 (1979).

28

uncertainty over whether the pursuit of conduct aiding the interests of the principal will actually increase

the rewards given to the agent.

In addition to different accuracy in their perceptions of desirable results, different biases of

principals and agents in measurements of results may cause the conduct of an agent to diverge from that

which will best serve a principal. Different biases in information measurements may arise for a number

of reasons. For example, a principal may be better able to perceive increases in the desirable features of

results of an activity than the agent who produces the results (perhaps because the principal is the

immediate recipient of the benefits). Alternatively, an agent may be able to perceive the costs and risks of

particular conduct more accurately than a principal (perhaps because the agent is carrying out the conduct

and bearing associated costs and risks as he or she produces results). In either of these types of cases,

increased risk to the agent will result. Either of these asymmetries will cause a principal to reward to

encourage and promote conduct by an agent which the agent knows as not being the best way (in light of

full information on costs and risks) to promote the best interests of the principal.

b. Risk Averse Agents

If an agent is risk averse, the agent will tend not to take optimal actions to pursue his principal's

interests if the agent's incentives are tied just to the attainment of results desired by the principal. The

agent's risk aversion will cause the agent to fail to pursue certain risky actions that would, on balance, be

projected to promote the principal's interests, but which are rejected by the agent due to the latter's risk

aversion. Hence, a simple fee structure that directly links improved rewards for an agent to increases in

the value of actions to the agent's principal will not cause the agent to balance risk, projected return, and

action in a way that will optimally promote the principal's interests. Instead, certain actions falling

between the risk preferences of the principal and the agent will go unaddressed.

To offset this problem, the optimal fee schedule linking the interests of a risk averse agent and her

principal will sometimes need to offset the agent's risk aversion and encourage the agent to undertake

efforts optimally promoting the principal's interests despite the agent's concern over particular risks.

Specifically, a risk averse agent must sometimes be paid a fee associated with efforts tending to support a

successful result regardless of whether the efforts produce such a result.

This type of fee schedule encourages the agent to take actions that entail risks that the agent

would not personally be comfortable in bearing, but that are nonetheless associated with optimal pursuit

of the principal's interests. At the same time, by tying a fee to the completion of efforts promoting

success, the agent encouraged to pursue those efforts. Where some or all of the rewards given an agent

are linked to efforts in this way rather than to results achieved, the principal is in effect taking the risk that

the efforts generating the reward will not produce any benefit to the principal. This element of risk has

been shifted from the agent to the principal, thereby lessening the risks born by the agent.

The fee schedule that will produce optimal conduct on the part of a risk averse agent is a schedule

that shifts a sufficient quantum or feature of risk from the agent to the principal to make the agent risk

neutral as to the residual risk that the agent still bears. Where there are certain types of risks that an agent

is particularly averse towards and other risks that the agent tends to be risk neutral towards, effort-link

elements of a fee schedule may be tailored to shift to the principal the types of risks that the agent is most

concerned about and risk averse towards. In other areas where the agent is risk neutral, successful results

should remain the scaling criteria for fees paid to the agent.

Likewise, even if there is only a single dimension of risk faced by an agent and the approach to

separating risks just described is not an option, then risk adjustment can still be achieved through effort-

related rewards to shift risk taking by an agent to a neutral point. This can be achieved by shifting enough

29

risk from the agent to the principal to make the relevant agent risk neutral regarding the remaining risk he

or she faces, but with no more shifting of risk as this would make the agent excessively risk preferring.

Once sufficient risk of failure has been placed on the principal to make the agent risk neutral regarding

further marginal risk, remaining components of fee rewards should depend solely on the attainment of

successful results.

D. Using Agency Theory to Interpret Non-Contractual Agency Relationships

Most agency theory analyses address contract-based agency relationships such as the sales

employee example described above.48

However, to evaluate patent-influenced innovation, agency theory

must be extended to assess the types of non-contractual innovation relationships implemented via patent

rights and incentives. Given that agency theory has seldom been used to assess non-contractual

relationships between principals and agents, some examples of how agency theory can be used to interpret

non-contractual relationship may help to suggest how agency theory can be extended to patent-mediated

relationships.

Steven Shavell has used agency theory to interpret varying allocations of incentives and risks

within tort systems applying negligence or strict liability schemes of safety-promoting liability.49

He

considered the implications of tort liability incentives in creating agency relationships in situations where

a firm is a potential source of accidents having adverse effects on others (such as environmental damage)

and where the probability or severity of the accidents is influenced by the actions of the firm. Liability

standards may encourage a firm in this situation to act as the agent of a potential victim of accidents by

taking safety measures that help to protect the potential victim. Government action may encourage safety

measures in these situations by imposing either negligence liability (triggered only where an acting party

fails to exercise due care to prevent harmful accidents) or strict liability (imposed in every case where

there is an accident). According to Shavell:

"The choice between strict liability and negligence may be thought of as a choice of fee schedule

for a principal (the government or society) and an agent (a firm). In this view, under strict

liability an agent's fee (zero of there is no accident and negative otherwise) depends only on the

outcome (accident losses), whereas under a negligence standard the fee depends on information

about the agent's actions."50

Shavell concludes that, so long as an actor is risk neutral in these situations, a system of strict liability will

tend to encourage the actor to take an optimal level of safety measures to protect society. However,

where an actor is risk averse, a negligence standard in which liability is sometimes reduced based on the

quality of efforts of the agent will sometimes be preferable to a strict liability scheme which would cause

the actor to invest too much time and resources in safety promoting efforts.51

Similar non-contractual agency relationships arise out of government-imposed penalties for

pollution. Fines for illegal pollution impose a payment or fee schedule on potential polluters that

encourages them to act as agents of the public in seeking to prevent harmful pollution. Where the

48

See text at Section III (B), supra.

49 Steven Shavell, Risk Sharing and Incentives in the Principal and Agent Relationship, 10 Bell J. of Econ.

55, 65-66 (1979).

50 Id. at 65.

51 See id.

30

imposition (or size) of pollution fines depends on a showing of harm to community members from

pollution, the threat of pollution fines sets up a contingent penalty system in which parties engaging in

harmful conduct (the pollution) are threatened with harmful consequences (the fines). This type of

contingent penalty system is the inverse of the sort of contingent reward system associated with most

agency relationships, but acts in the same way as a contingent reward system in causing agents to value

and pursue the interests of principals. Contingent penalties in environmental laws tend to link the

interests of potential pollution sources and potential victims. The fines that link the interests of these

parties ensure that pollution control measures which aid the public (by reducing losses from pollution)

also aid the firms (by avoiding the negative economic impacts of large fines).52

E. Interpreting Innovation Processes Using Agency Theory

Agency theory is a useful tool for analyzing innovation efforts because many innovation projects

involve efforts by innovation agents (inventors) who design and refine inventions to serve the interests of

other principals (that is, the parties who are the intended users of the inventions). Of course, sometimes

innovators invent advance to serve their own interests and then they or others share the innovations with

additional parties. While it would seem that these situations involve solo efforts, this would be deceiving.

Agency processes still have roles in these settings. Agency processes may influence how innovations are

expanded or refined as parties who develop advances to solve their own specific problems realize that

their solutions have broader implications and seek to generalize their advances to serve a wider audience.

These efforts to redesign and generalize an invention are driven by agency processes even if the initial

design efforts that produced the root invention were individual efforts to serve personal ends.

Even in their initial innovation efforts to serve their own interests, agency forces may encourage

self-interested innovators to think broadly about whether (and in what forms) their individual problems

may be held by others and to tailor inventive projects from the outset to target a broader user group. If an

innovator begins to appreciate that a particular innovation, if successful, will be useful to many users in

addition to the innovator, the potential benefits to these further users may spur more extensive or different

development efforts than if an innovator were only trying to produce an item or practice that is of use to

him or his company. The needs of broader users will establish a different perspective on innovation with

different inventive goals than just the objectives and needs of the inventor standing alone. Both the type

and scope of design efforts that an innovator will invest in a design project may be influenced by the

concerns and interests of other invention users. Attention to the interests of these other users will make

the innovator an agent of the additional users, at least in part.

The agency relationships in play in these situations are especially complex because the agent

must define many of the key terms and features of the agency relationship. Problem definition and

technology assessment are key parts of agents’ tasks in innovation processes. Principals cannot define

precisely at the outset of these processes what they want agents to deliver in the form of inventions.

Rather, principals (or the demands created by market forces representing the desires of principals) define

functional needs and desires which establish the work content and reward structure of resulting agency

relationships aimed at innovation. Part of the work of an innovator acting as an agent in this setting is to

define more fully the needs of the principals and then to design means to fill these needs. Both the

diagnosis and definition of needs and the creation of technology designs for needs fulfillment are

important agency tasks in innovation processes. This type of diagnosis of principal needs by agents is not

unique to innovation processes, however. It is a common feature of agency relationships in which expert

agents serve the needs of less expert principals. For example, when a patient (the principal) goes to a

52

See id. at 55 n.2.

31

dentist’s office, the dentist (the agent) both evaluations what dental work the patient needs (a diagnosis

and needs definition step) and then completes the needed work (a solution delivery step).

Framed in terms of agency relationship dynamics, many aspects of innovation processes can be

interpreted and potentially improved. For example, key innovation features such as the risks allocated to

innovators and innovation users and the levels of rewards needed to encourage efficient levels of

innovative efforts by potential inventors can be modeled in terms of agency theory. These sorts of

agency-based models are applied in the remainder of this article to reveal, formalize, and interpret various

features and implications of patent impacts on innovation processes.

IV. Agency Frameworks for Evaluating Patent-Influenced Innovation

A. Factors Influencing Patent-Mediated Innovation

The formal description of agency incentives described in the previous section provides us with a

structure for interpreting the incentives for innovation created by patent rewards and the factors that vary

the influence of those incentives. Viewing the aggregate user set for an invention as the principal on

behalf of whom a party working on the innovation is serving as an agent, the expected aggregate utility of

the invention to users is:53

Expected Aggregate User Utility = ∫ ∫

where:

U(x) = users' utility Function

e = agent's effort

z = users' observation of e

x = outcome

f(x,z) = fee paid by the users to the agent (a function of x (outcome) and z (observation of effort))

r(x|e) = probability density of x given e

The agent's perceived utility in this situation is given by:

Expected Innovator's Utility = ∫ ∫

where:

V(x) = agent's utility function

q(z|x;e) = probability density of z given x and e

53

This analysis assumes that there are no prior items or practices that address the user needs served by the

new innovation. If such prior items or practices exist, then the analysis presented here should be adjusted

to focus not on the total utility achieved, but rather just the net utility over and above the net utility

achieved by using the preexisting technology. This is equivalent to saying that the origin from which

utility is measure in the text should be adjusted to equal the starting point when the new innovation is

developed and introduced. This starting point may include existing technological tools addressing the same

practical problem as the new innovation, in which case the analysis here should focus on functionality

improvements and increased utility from that starting point.

32

While these formulas do not dictate the precise utility functions that will influence specific

innovation projects (which will be subject to context-specific consumer demands and product production

challenges), these formulas do reveal the types of factors that will vary the impacts of patents across

various situations and types of innovation projects. These formulas remind us that innovation under

patent-mediation will depend upon:

1) Fee Amount Paid for Innovation Success: The greater the fee paid to an innovator for an

innovation result x, the greater an innovator’s motivation will generally be to create that result.

This factor (represented by the dependence of the fee size factor f(x,z) on results x in the above

model) means that fee structures providing greater rewards to innovators for successful

innovation results (roughly corresponding to substantial innovation value delivered to innovation

users) will typically produce more motivation and more intensive innovation efforts than smaller

rewards for the same results;

2) Accuracy of Observations of Success: For a given innovation result x, the more accurately that x

can be perceived and the intended reward for x paid, the more clear an innovator’s motivation

will be to create result x. This factor (reflected in the dependence of f(x,z) on observational

quality z in the above model) means that where an innovator must worry about observational gaps

such that desirable innovation success may be attained without the innovator being paid the

normal fee for producing that result, the innovator’s motivation to produce the result will be

discounted and decreased accordingly. The offshoot in cases of high observational error or

difficulty will tend to be a reduction in related innovation incentives and efforts;

3) Effort Costs Reducing Net Fee Value: The costs (or distastefulness) to agents of rendering

various innovation efforts will influence the value of these efforts to agents and potentially reduce

the net value of innovation incentives to such agents. This factor (reflected in the dependence of

V(f(x,z),e) on e as well as fee size f(x,z) in the above mode) means that innovation process costs

to agents (including opportunity costs) will be constant drags on innovation systems, leading to

reductions in innovation settings where such costs are large. Also, since these costs are as

measured by the agents who are innovators not the parties receiving the benefits of innovations,

the costs limiting innovation results may not be easily perceived by innovation users leading to

mysterious barriers to innovation progress;

4) Influence of Effort Variations on the Observation of Innovation Success: The extent to which

different efforts by an agent can enhance the likelihood that successful innovation results will be

recognized and rewarded as such will also influence the size of projected rewards perceived by

agents and the scope of resulting innovation incentives. This factor (reflected in the dependence

of innovation rewards on q(z|x;e) in the above model) indicates that agents may tend to choose

efforts that either enhance the accurate observation of desirable innovation outcomes (perhaps by

participating in related technology publicity or explanations for consumers) or may seek (if

possible) to fabricate false evidence of innovation success to the extent that this will create the

appearance of project success and produce rewards for that success. Provided that mechanisms

to reveal and prevent such fraud are in place, generally innovators will have a stake in shaping

their efforts to enhance the likelihood that innovation users will perceive the value of new

innovations since this will enhance the likelihood that innovation rewards based on the user value

of a new advance will be paid;

5) Probability-Weighted Aggregation of Rewards Across Innovation Observations: Since agents

cannot count on a particular level of observation of their innovation results by innovation users,

such agents cannot project their probable innovation rewards based on a particular observation

level. Rather, these agents must estimate their likely rewards from involvement in an innovation

33

project across a spectrum of different observation levels and resulting reward variations. This

factor estimate (reflected in the above model in the integration of probability-weighted value

estimates V(f(x,z),e)q(z|x;e) across all possible observation levels z) means that the most likely

clarity levels in the perception of project outcomes will dominate the perceived value of

innovation efforts by potential innovators. If most project outcomes are poorly perceived by

potential innovation users, then many successes will probably be overlooked and large project

rewards to innovators will be correspondingly rare. If most project outcomes are clearly observed

and rewards can be based mostly on these clear observations, then the net projected size of

innovation rewards will tend to be higher. The aggregation of the perceived value of innovation

rewards across the probability of various levels of observations recognizes that different product

perfection and marketing programs may translate particular innovation successes into highly

different presentations and evaluations of innovation results by potential innovation users with

highly different reciprocal payments of innovation rewards to the innovators involved. The

integration of potential rewards across various observation levels provides agents with a means to

consider the spectrum of possible observation processes and to create a single blended view of the

value of innovation project involvement for an agent in light of the later spectrum of innovation

observation accuracy.

6) Influence of Effort Variations on Innovation Success: The extent to which different innovation

efforts by an agent can enhance the likelihood that successful innovation results will be realized

will also influence the size of projected rewards perceived by agents. This factor (reflected in the

dependence of innovation rewards on r(x;e) in the above model) indicates that choices by agents

about the scope and nature of efforts to exert in pursuit of innovation will enhance invention

rewards where increased efforts (or different types of efforts) enhance the likelihood of successful

innovation. If this is not the case – perhaps because an innovation project is technologically

impossible no matter how much effort is invested or because the innovator in question plays a

small role in a project and cannot significantly influence the outcome no matter how much effort

is expended – then the innovator involved will not be incentivized to change his or her effort

levels to achieve invention rewards;

7) Probability-Weighted Aggregation of Rewards Across Possible Outcomes: Potential agents

contemplating participation in innovation projects will not perceive a single possible reward

outcome, but rather will realize that various innovation project outcomes are possible, each with a

different possible reward to the agents. However, this will not preclude a single perceived

incentive for a given project – rather, the perceived reward for involvement in the project will be

a probability-weighted sum of the various component rewards for different project outcomes.

The likelihood of receiving any particular reward will be discounted by the fractional likelihood

of getting that reward. The overall projected value to an agent of being involved in a particular

project will be the sum of these probability-weighted reward estimates. Through is summing

process, an agent can project the most likely net value of project involvement to him or her taking

into account the full range of project outcomes that may influence the agent’s rewards from the

project. This aggregated reward estimate (reflected in the above model in the integration of

probability-weighted value estimates V(x) across all possible project outcomes x) means that the

most likely project outcomes may dominate the reward values perceived by innovators. It may

also mean that if innovators feel that they can significantly enhance the likelihood of project

success (thereby increasing the weighting of high rewards related to project success), the agents

may see significant value and attraction in being part of an innovation project; and

8) Innovator’s Utility Function: The receptiveness to rewards presented and the risk preferences

potentially discounting this receptiveness will dictate the net value of patent-based rewards

perceived by an innovator (V(x) in the above model). This means that the type of rewards valued

34

by an innovator (and the degree of value seen in both patent-based rewards and other rewards for

alternative conduct) will influence the motivating impacts of promises of patent-based rewards

and affect the resulting scope of innovation efforts.

These agency innovation factors suggest some of the patent system leavers that both individual

innovation users and public policy makers can use to increase the impacts of patent-mediated innovation

and the public value of resulting innovations. By varying these types of information and fee factors,

patent-based processes can more clearly link perceived innovation value as seen by innovation users with

corresponding patent rewards that incentivize specialized parties to pursue innovation solutions and create

more useful technologies.

B. Optimal Patent-Mediated Agency Relationships Promoting Innovation

`

Under the incentive and reward structure just described, optimal action to promote the interests of

the user-principals will be encouraged if an incentive fee scheme f is presented to an innovating agent and

the scheme is Pareto optimal in the following sense:

1) The expected aggregate utility of an advance to invention users is maximized over

alternative fee schemes;

2) For a given fee structure, the innovator's utility is maximized over alternative effort

levels; and

3) The innovator's utility is at least equal to or above the utility associated with alternative

conduct.

Each of these three considerations deserves some further attention.

1. Implications for Innovation Users

The first of these considerations suggests that while various fee structures (as implemented by

contractual payments promised by users or by rewards achieved by various versions of patent rights) are

possible, fee structures that produce large user benefits leading to large aggregate user utility will

generally be desirable over alternative fee schemes. Higher fees will tend to increase aggregate user

utility where the fees encourage more effective innovation efforts54

resulting in more utility per resulting

product unit. Higher fees may also increase aggregate utility where the fees encourage innovators to

produce designs that serve the interests of more numerous invention users (since more users will tend to

realize more utility than fewer users). Where higher fees do not tend to produce more effective products

(from the perspective of aggregate utility achieved), a fee scheme incorporating these fees will be

undesirable. This will be true because the net utility to users (the overall aggregate utility less the higher

fees paid) will be smaller than the net utility from a scheme charging lower fees but achieving similar

innovation benefits for users.

Blends of fees paid for efforts and results may be needed to optimally incentivize desirable

innovation efforts, particularly where the bulk of potential innovators are risk averse. As previously

discussed, if agents pursuing innovation are risk averse, a fee structure directly tying increased fees to

increased utility gains realized for innovation users may not produce optimal innovative efforts. Agents’

risk aversion will cause them to fear to embark on certain innovation projects with significant risks of

54

Increased fees may encourage this by either incentivizing innovators to work more on their advances to

improve them to higher utility levels or causing innovators to focus more effectively on product utility in

setting product design criteria and producing more useful product designs.

35

failure even though a risk neutral agent would pursue the same project and the relevant principal would

benefit from the project. Risk averse agents will balk at certain risky types of innovation projects that

exceed their limited risk preferences. In these situations, an optimal fee schedule will shift some of the

risk of failure of an innovation project to the principal involved (for example, by paying innovators

straight salaries for innovation efforts and further bonuses or other incremental compensation for

successful efforts providing increases in benefits to innovation users.

In addition to helping to offset the risk aversion of some agents acting as innovators, a fee scheme

that turns (at least in part) on efforts rather than results may be easier to implement than a scheme turning

on results because the relevant efforts are easier to observe and interpret than the results. Information

gathering costs may also influence choices about optimal reward criteria. To the extent that observations

are imperfect (or that improved observations are highly expensive to obtain), it may be highly expensive

to administer a fee payment system that accurately monitors agents’ actions and only grants rewards

where desired efforts are undertaken or results achieved. Measurement costs in administering innovation

fees and rewards may be large enough to be their own compelling design limitation. The size of

measurement costs may overshadow the advantages of gaining better information and administering

innovation fees in a manner that more carefully matches success in innovation efforts or results. In these

situations, lesser emphasis might be given in a fee scheme to aspects of efforts or results that are

particularly difficult to monitor. For example, if the value of innovation results is difficult to measure at

the time a payoff to an agent is expected (perhaps because related products are still in development or

because the full range of users and user impacts of an advance are still uncertain), then a fee scheme

emphasizing the completion of efforts tending to produce useful results might be preferable to a scheme

in which the size of a reward for innovation turns on the usefulness of the advance to the public.

2. Implications for Agents Engaging in Innovation

Once a probable fee structure rewarding innovation is known, agents capable of completing

innovation projects will tend to jockey their efforts to maximize their own returns within the spectrum of

available rewards. This positioning behavior has several consequences under the proposed agency model

for innovation projects.

First, agents will tend to focus primarily on efforts or results to the extent that these factors

dominate the criteria for determining and increasing the agents' potential rewards. Where the primary

criteria for determining the rewards for innovation to be paid to agents is the aggregate increased value of

innovation results for innovation users, agents attempting innovation will have reasons to tailor their

innovation designs and their measures of innovation progress to the attainment of benefits for innovation

users. The overall practical value of innovation results to innovation users will be an indicator of rewards

payable for successful innovation results. By contrast, rewards which are dominated by higher payments

for completion of certain innovation efforts (hopefully, efforts tending to increase the chances of

successful innovation) will tend to shift agents' behavior towards the completion of those efforts

regardless of whether the efforts actually increase invention users’ benefits from an innovation or

otherwise promote the interests of innovation users.

Second, agents pursuing innovation may feel that many factors governing the success of

innovation results (particularly the popular distribution and commercial success of such results) will

depend on factors far beyond their control such that rewards derived only from that success and measured

solely from the commercial value of innovation will entail very high risks to the innovators. In these

circumstances, promised of rewards based on other factors than innovation results as measured from

innovation value to users may be needed to lower innovators’ perceived risks and create attractive

incentives for innovation efforts. This risk lowering can be achieved by rewarding innovation efforts

directly, either as a sole reward for actions supporting innovation projects (as where an employee is paid a

36

salary, but has no stake in the commercial success of a project) or with a combination of rewards for

efforts and results (as where an employee is paid a salary for innovation efforts and promised a further

bonus measured from the commercial success of project results). Such a blended set of incentives

combining low-risk rewards for innovation efforts and higher-risk rewards for innovation results may be

needed to motivate user-focused innovation activities by highly skilled agents who are capable of

technological innovation, but who are unwilling to bear the full risk that these efforts may fail to produce

useful, valuable results.

3. The Competing Lure of Alternative Activities and Rewards

The third necessary feature of an optimal innovation fee scheme is that the scheme should

provide an expected net reward to an agent that is at least as large as the reward the agent can gain

through alternative activities. This implies an important limitation on viable fee arrangements. Unless,

an agent considering whether to embark on an innovation project can project a probable reward that is

greater than the projected rewards available for other, less risky activities, then the agent will generally

not pursue the innovation project. Innovation in these circumstances will be trumped by the lure of

rewards from alternative activities. Because the time and talents of potential innovators are limited

resources – and in the case of our best innovators, highly limited resources – this tension between

innovation and potentially rewarding alternative activities is a constant threat to innovation efforts.

The background lure of alternative incentives and rewards dictates some important features of

agency relationships promoting innovation. First, background labor market conditions and career

alternatives for potential innovators determine what sorts of innovation arrangements and projects are

attractive. Where the expected payoff of an innovation project or activity is not demonstrably greater

than the pay an engineer or other innovator can obtain in more mundane but more certainly rewarding

lines of work, there would seem little economic reason for the engineer to pursue the innovation project

over the more certain source of income. Admittedly, other incentives (such as rewards from reputation or

prestige gains following successful completion of important inventions) may encourage innovators to

pursue projects where monetary rewards alone would not be sufficient to produce innovative actions.

However, even these additional factors and sources of motivation may only be sufficient to overcome

small gaps between projected economic rewards for high- and low-risk activities available to potential

innovators. Reputation and prestige gains from research success are potentially important motivators for

innovation effort, but may often still be trumped by the more certain projected economic rewards

available to talented engineers and scientists by foregoing high-risk invention efforts and pursuing more

certainly compensated alternatives.

Second, the lure of rewards for alternative activities means that highly capable individuals may

need especially large incentives to induce them to turn towards innovation projects amidst the many

highly compensated projects available to them. For the most talented members of engineering or

academic elites having state-of-the-art knowledge or training that enables them to produce exceptional,

non-obvious advances in their fields – the types of advances that patent laws and incentives are aimed at

increasing55

-- the background or "reserve" incentives needed to attract individuals to innovation projects

may be particularly large. These rare, highly talented individuals will frequently have significant value to

corporations, universities, or other employers in many capacities, providing many attractive alternatives

to these individuals in private or academic employment.56

The rewards available for these alternatives

55

See 35 U.S.C §103.

56 By contrast, the average engineer in a particular field -- a person with widely held skills who is capable of

altering or expanding past designs in the same field by applying analytic techniques that are used in day to

day activities in the field of interest but who will rarely go beyond this to realize non-obvious advances –

37

establishes a threshold level of compensation above which innovation project incentives must rise to

attract the attention of these essential individuals. Invention incentives must not only encourage invention

efforts, they must trump the many attractions and types of potential compensation that highly capable

(and often highly compensated) individuals will have.

C. The Model Illustrated: A Simple Agency Innovative Process

To illustrate some of the analytic strengths of the agency model of innovation and incentives just

described, this section describes the agency features embedded in a simple innovation project. In this

project, a single agent works on an innovation on behalf of a particular principal.57

Assume that a

particular business person A operating a retail store engaged in widget58

sales (the principal) wishes to

improve the widget containers used in his retail business. Widgets are fragile devices that are known to

become inoperative unless kept in special containers that maintain carefully controlled surrounding

temperature and humidity. A is dissatisfied with the widget containers currently used in his retail stores

and contracts with engineer B to produce an improved widget container. The contract specifies that a

successful container design must reduce spoilage levels or electrical operating costs (or both) relative to

the widget containers presently used in A’s retail stores.

The contract between A and B specifies that B's efforts will last for three months, at the end of

which B will deliver her new container design. At that time, if she is successful in producing an

improved design (as measured under the above design criteria included in the contract), B will receive a

design fee. In negotiating their contract, A and B initially present two different ideas about how this fee

should be determined. B wishes to receive a flat fee for each hour she spends on engineering efforts

related to the contract. She is not sure that she can produce a design with the features that A is looking

for, but B thinks that there is some hope of success. Since she foresees a substantial chance of failure, B

is not willing to contribute her own time and effort to the project at the risk of losing these resources if

she cannot produce a successful widget container design with the specified features. B, in effect, wants to

put the risk of project failure on A. B points out that she has recently been employed at $50 per hour as

an engineer and expects to gain at least this level of income from the present project.

A, on the other hand, hopes that the contract will provide for no payment until B produces a

successful design and then only a payment that is scaled to the cost savings achieved by the new device.

A proposes that B's fee equal the cost savings realized by A during a year of using widget containers

will be more numerous and more easily drawn into innovation projects. For such an individual with widely

held skills, the reserve level of rewards that must be overcome by innovation incentives to encourage

innovative risk taking and participation in innovation efforts is lower than will be the case for those few

innovators with exceptional, highly compensated skills. Elevated incentives -- perhaps administered

through patent rights -- may need to be offered to scientists and engineers with rare, highly compensated

knowledge or skills to lure these parties with "non-obvious" insights into the innovation game as agents

serving potential innovation users. At the same time, equivalent incentives may not be needed to attract the

more common engineers with average skills. This distinction in reward levels is currently observed within

patent laws by offering patent rewards to parties producing non-obvious inventions but precluding patent

rights for other new types of advances that are mere obvious extensions of previous designs in the same

field.

57 See Stanley Baiman, Agency Research in Managerial Accounting: A Survey, 1 J. of Accounting Literature

154, 164-65 (1980) (presenting a similar example of a simple principal and agent relationship in the context

of a product maintenance project).

58 The widgets involved here are characteristic of widgets generally. They are presumed to be hypothetical

items having useful (if mysterious) qualities, the nature of which is not germane to our discussions.

38

based on the new design. For this year, all of the net benefit of the new design will go to B, creating

strong incentives for B to pursue and increase these benefits. After the first year, all of the same benefits

will go to A, providing incentives for A to pursue the project.

After some negotiation, the parties agree on a fee structure that includes both efforts-based

rewards (in the form of a payment of $30 per hour for services rendered) to interest B in starting the

project and a further results-based payment for successful delivery to A of a widget container meeting A’s

design criteria. The payment for successful project completion is set at 50 percent of the cost savings

realized by A over one year of use of the widget container delivered by B (thus creating a reward for

achieving desirable results for A and for increasing the scope of benefits transferred to A via a successful

design). The project completion payment is to be paid a year after B delivers a widget container

reflecting A’s design criteria. The blended combination of efforts-based compensation and results based

additions both reduces B’s risk in the project and still somewhat links her incentives and innovation focus

to the functional results and value desired by A.

In carrying out this contract, A and B will need to complete three monitoring steps and interpret

three corresponding performance signals. First, in qualifying for her hourly fee, B will need to

demonstrate reliably that she has spent the hours of work she claims on efforts related to the innovation

project. Second, at the final delivery stage the design submitted by B (if any is submitted) must be

measured against the criteria for success specified in the contract. Third, assuming that it is a successful

design, the further cost savings realized by A must also be measured to determine the size of the reward to

B that is due.

Several types of measurement problems may frustrate the administration of this agency

arrangement to ensure payment of rewards to the agent in ways that will best promote the interests of the

principal. Rewards for efforts should ideally be paid only for efforts that promote successful results for

the principal, yet limiting reward payments to these circumstances entails two potential measurement

problems: first, potential inaccuracy as A seeks to measure the scope and quality of B's efforts to modify

A's old container designs and, second, even if A’s efforts can be measured with perfect accuracy,

potential inaccuracy in evaluating whether those efforts actually increase the likelihood of a successful

project outcome. Inaccuracies in either of these types of precursors to reward payments may result in

payments made for dysfunctional actions by B and inefficiencies in the innovation incentives aimed at

promoting delivery of a successful design to A.

Other measurement problems and uncertainties will add inefficiencies at the point where B’s

delivered designs are assessed as a possible basis for payments of rewards to B for successful project

completion. While the use for a year of a new widget container conforming to B's design may provide a

practical measure of the probable value of the design to A, the representativeness of this measurement (or

A's honest reporting of the scope of the benefits he achieves) may still be in doubt. Measurements of this

benefit will probably be somewhat uncertain and B will be correspondingly uncertain about the scope of

the results-based rewards she can expect. In the face of such uncertainty, B may cut back on her design

efforts to a level that is below the level that would have produced more beneficial design results from A's

cost/benefit perspective.

In light of these sorts of measurement problems and resulting inefficiencies, A's challenge as

principal is to choose B's payment schedule and the associated monitoring steps for gauging B's efforts

and results so as to maximize A's own expected utility while inducing B to act as an efficient and

effective agent advancing A's interests. B's aim as an agent is to supply the type and amount of design

effort that will maximize her own expected utility given the agency relationship’s payment schedule and

monitoring system. In determining the desirability of any payment schedule-monitoring system

39

combination, A, as principal, must analyze what level of innovative effort B, an agent pursuing his own

best interest, is likely to supply given the incentives of the payment schedule-monitoring combination..

D. Extending the Agency Innovation Model to Multiple Agents and Multiple Principals

Innovation opportunities are often pursued simultaneously by multiple agents (in the form of

competing innovators) serving multiple principals (in the form of groups of potential users of a single

innovation). The involvement of multiple parties on either or both sides of principal-agent relationships

for innovation may affect the formation, execution, and effectiveness of these agency relationships. The

involvement of multiple parties may add a number of complicating features to agency processes

supporting innovation, including enhanced transactional costs in administering these processes and

expanded opportunities for strategic behavior. This subsection explores some of agency implications

associated with multi-party agency relationships for innovation.

1. Implications of Parallel Innovation by Multiple Agents

Where multiple parties work in parallel to produce an advance, the nature of the additional

agency process features will depend on whether the multiple agents pursue innovation in a cooperative,

independent, or strategically negative manner. Each of these possibilities is considered here with

attention to the impacts of these behaviors on optimal fee schemes for incentivizing the multiple agents

involved.

a. Independent Agents in Competition

Where a single user would benefit from a particular innovation, the user may wish to promote

competition to produce the innovation among multiple agents. For example, in the context of the

container project just described, the potential innovation user A many know of two engineers B and C

who appear capable of developing a new design for an improved container. A might establish a

development contract with each of these parties in which the first of B or C to submit a satisfactory

container would be due a fee, with the amount of the fee to be determined based on the amount of savings

achieved by A as already discussed. In such an arrangement, the slower party would receive nothing.

Such an arrangement would establish the counterpart to bounty arrangement in the Old West in which the

first party to deliver a specified criminal to public authorities was promised a bounty payment. These

types of historical bounty payments rewarded speedy and effective work since the first to bring in a

criminal received a reward while others who may have been looking for the same criminal but who were

slower or less effective in catching and bringing in the criminal received nothing. In the same way, a

bounty contract related to innovation can place a premium on speed and effectiveness in design

generation, with the first submitting a successful design qualifying for a payment and the slower or less

effective designers attempting similar innovation receiving nothing.

Where multiple agents are aware of the competing efforts of other agents in a situation like this,59

but the agents are unable to interact to form coalitions so as to cooperate in their innovation efforts, the

innovation incentives seen by each agent will be discounted by the risk that their efforts will be trumped

59

Where multiple competing agents working on an innovation project are not aware of each other, the

incentives perceived by each may be very similar to those perceived by a single agent working on a similar

project. However, the offering of an innovation reward under terms that indicate the reward will be paid to

the first party who produces a particular result implicitly raises the possibility that there may be multiple

parties pursuing this reward. The reference to first submission suggests that there may be second and

subsequent submitters who will have worked on the project yet receive nothing.

40

by a competitor and they will receive nothing. Each of the innovators will feel some uncertainty about

the progress and success of the others, adding a source of perceived risk that the individual innovator will

be beaten out by a rival and will both not qualify for the reward at issue and need to bear any costs they

have incurred in their unsuccessful innovation attempts. Information gained about the relative

competence and progress of competing innovators will affect the size of each innovator’s incentive

discount due to fears of competition.

If a particular innovator has sufficient private information to believe that he need not worry much

about competition by rival innovators, the discounting effect due to fear of rival innovators may be small.

This may be the case, for example, if an innovator, although aware of competitors, feels that she has

significant background knowledge in an application area or insights about the capability of a technology

not shared by her rivals. However, if the rivals seeking innovation simultaneously are both aware of each

other and consider themselves as relatively equals in the competition to produce a successful advance, the

discounting effects of this awareness of competition may be considerable. A potential innovator

considering an innovation project under these circumstances may significantly curtail the efforts he or she

is willing to contribute to a particular project or may choose to turn away from participation in a particular

project altogether in favor of pursuing alternative efforts and sources of income with more less

uncertainty. Concern over competitors may also bias innovators in favor of quick innovation methods

over slower but arguably more effective methods as the quick methods will be most likely to help

innovators win a “race” for innovation rewards that are available only to the first producer of an

innovation.

Where incentives to potential agents are somewhat weakened by fears of rivals, principals

seeking innovation can somewhat overcome this problem by establishing additional fees promoting entry

into innovation projects. Where the risk of losing to a rival innovator is real, each innovator may refuse

to embark on an innovation project unless paid a substantial fee for their efforts. This type of payment

would overcome the innovators' concerns over losing out to rivals, but would also add to the costs of

innovation for the principal or the amounts that the principal can pay for results achieved in the

innovation project.

b. Independent Agents Hindering Rivals

Where multiple independent agents pursue innovation in parallel and are both aware of their

rivals and have opportunities to influence the activities of their rivals, the multiple agents may interact in

negative ways to impede the chances of success by rivals. Particular agents may see an advantage in

impeding rivals in this way if such negative actions are either undetectable or at least not penalized within

the applicable reward scheme incentivizing innovation. Where the primary rewards for innovation (such

as patent rights and associated commercial returns) primarily involve payment of only one fee for the first

to produce a particular innovation, impeding efforts need not completely interfere with a rival’s efforts,

but need only slow those efforts down enough to insure that the impeding party wins the race to be the

first innovator and thereby qualify for the available innovation reward. The threat of innovators’

hindering efforts will reduce the prospects of success perceived by particular competing innovators,

thereby reducing the net incentives to pursue innovation projects and the range of advances those projects

will produce.

c. Cooperative Multi-Agent Actions

Where multiple agents can bind themselves together and jointly pursue an innovation project

through coordinated actions, a number of benefits may follow and many of the potential problems of

rivalrous action can be avoided. Multiple innovators working together can produce an efficient team

effort in pursuit of a single innovation reward to the group (which they can agree among themselves to be

41

split in a pre-determined manner). The group members will not have to worry about other group

members as rivals (assuming that none default from their group arrangements and commitments). The

group will see the full value of an offered reward as the incentive for pursuing an innovation project.

They will not need to discount the predicted value of this reward due to uncertainty about whether they

are going to be beaten out by rivals (except, of course, further rivals, if any, who are not part of their

innovation group).

Coordinated efforts among team members can heighten the efficiency and effectiveness of

innovation efforts by avoiding duplications of efforts, realizing economies of scale, and allocating

specialized tasks to team members with the best specialized knowledge and skills to perform those tasks.

Some projects may require contributions by more than one innovator, either because contributions from

very different fields are needed to complete the projects or because the tasks needed for completion are

too numerous for one person to complete in a timely manner. Team innovation will be the only way to

compete innovation projects in these sorts of circumstances. By realizing the various advantages of

teamwork, multiple innovators working in a single team may produce more efficient and extensive

innovation results than the same number of innovators could have realized if acting as generalists in

isolation.

Of course, working in a group through cooperative action has its own costs. Innovators working

in a team -- perhaps in a joint venture or a startup company -- need to define, coordinate, and combine

their individual efforts to accumulate results that successfully complete an innovation project. They also

need to agree upon means for allocating rewards received for successful project completion (or at least

upon a mechanism for allocating such rewards if the rewards are received). The group members also

need to work out the ongoing administrative details of operating as a group and communicating about

partial progress and problems to other group members. These various coordination and operational tasks

will add substantial administrative costs to group operations over comparable individual projects. Intra-

group interactions among team members will also create opportunities for strategic behavior among group

members as some individuals with the group seek to maximize their returns at the expense of other group

members. The resulting administrative costs and strategic behaviors will tend to reduce the innovation

incentives seen by some or all of the group members, thereby diverting them from some of their potential

focus on the needs of principals in the relevant innovation projects.

2. Implications of Innovation Benefitting Multiple Users

The addition to the model of agency relationship for innovation of multiple potential users of an

innovation may add considerable complexity to the formation and operation of such relationships. This

type of agency relationship involves multiple principals who can either offer individual rewards for the

supply of advances serving their individual needs or who can band together to establish an overall reward

for an advance serving a common need. The first situation can be modeled as numerous separate agency

relationships that happen to be coexisting at the same time. Since the implications of agency innovation

serving single principals have already been considered, this situation involved parallel repetitions of these

individual-focused relationships will not be given additional attention here.

The possibility that multiple principals may act in concert is a new circumstance not yet

considered. This potential for group action on the principal side of agency relationships may add both

efficiencies and problems to innovation processes. To examine the implications of adding multiple

potential users to the agency innovation model, consider an extended version of the container design

project already discussed. Assume that instead of just A as a single potential user for the targeted new

container there are approximately 1000 retail stores with needs similar to those of A. The operators of all

of these additional stores might benefit from an improved widget container design. For simplicity,

assume that B is the sole innovator who seems capable of producing an improved design in an efficient,

42

effective manner. Although B's efforts could produce benefits for the multiple store operators, the

process of bringing the store operators together to offer and administer rewards for innovation (or of

finding a surrogate for these multiple operators in the form of a product developer and producer who will

take the development risk on behalf of the multiple container users) can add considerable complexity to

the innovation process and undercut related innovation incentives.

Free riders among the multiple store owners will be an important concern in this setting. If a

particular user C thinks that efforts she funds will work to the special benefit of another rival store owner

D, then C will be hesitant to fund and reward innovation efforts by B. C may be hesitant to contribute to

innovation incentives if she fears that D will not bear his fraction of the overall container development

costs (placing more of these costs on C and the other store operators who agree to contribute to the

development project). Where C and D are in direct competition, these fears will be justified. D will be

able to use the resulting container in business activities that will aid D’s retail store competition with C.

If D does not pay the same product development fee that C pays, D will eventually conduct business

activities that leave him stronger than C. From C's perspective, this scenario will seem like putting C's

money in the pocket of a rival, strengthening D over C in subsequent competition between the two.

Rather than providing D with this advantage as a "free rider" who gains B's efforts without paying a share

of the costs of those efforts, C might forgo involvement in the innovation project altogether. C may

rightly believe that moneys she would have used to incentivize B's innovation are better used to fund

business activities with a greater certainty of a profitable return to C and no benefit to D.

One way for C to avoid these sorts of free rider effects and to help justify C’s investment in B’s

innovation efforts is by limiting access to the results of B's work to C (and others like C who pay for and

incentivize B’s work). Limited access might be accomplished through contractually-imposed access

limitations and trade secret protections if the new widget container can be used in secret and trade secret

protections retained for the features of the new design. This strategy would lessen fears of free riders and

force competitors such as D to bear their own development costs if they wanted access to a similar

container design. However, this type of system of parallel container development raises its own

problems. First, under this scenario, multiple container users (or groups of users) will need to identify

and contract with parties capable of innovation efforts, leading to duplicate transactions and waste in

innovator identification, contract formation, and innovation activities. Furthermore, assuming that B is

the best qualified party to design the relevant innovation and his services were reserved to the first parties

who engaged him, subsequent contracting parties or groups would have to make due with lesser

innovators and less effective innovation projects. Finally, a trade secret strategy may simply be an

ineffective access control mechanism for a publicly disclosed product – like a product container – where

the key elements of the design will be revealed through commercial activity and related public

revelations. Once these revelations have occurred, there will be no trade secret protections left for the

revealed design details. Other parties will be free to use these details in their own product containers,

leading to the types of free rider consequences already discussed.

Putting free rider behavior aside, there are many other administrative difficulties and costs that

may frustrate the formation of effective incentive contracts for innovation involving large groups of

potential innovation users. To create a single contract that established innovation incentives on behalf of

all users and that avoided the types of duplications of efforts just discussed, the parties seeking to create

and administer an agency innovation process on behalf of numerous principals would have to overcome

several hurdles. The multiple users interested in a particular type of advance would need to be identified,

enter into a contract defining an agent’s innovation fee and a fee sharing mechanism, specify further

administrative arrangements governing the monitoring of the agent’s efforts and the payment of the

relevant fee upon a successful result, and further agree on a means to share ownership or control of rights

and access to the resulting innovation.

43

While agreement on such complex terms might be possible, at some point the size and complexity

of a potential user group would add such costs to the contracting process that full groups of innovation

users would not participate in creating innovation incentives for potential innovators. With some users

left out, the remaining users (assuming that they did band together in fractional subsets) would only create

innovation incentives that reflected the needs of this partial user group, thereby establishing smaller

incentives than if all the invention users’ needs and related incentives where applied to the innovation

project. The result will tend to be suboptimal incentives to innovating agents that encourage only those

innovation efforts that are reasonable in light of the faction of users participating, but that do not

encourage the broader set of innovations that would be reasonable in light of the full set of user interests

potentially served by the innovations. Inefficiencies and ineffectiveness in the contracting process will

tend to leave the interests of some users out of the incentive process and thereby create suboptimal

innovation incentives.

The problems of innovation group contracting become even worse if it is assumed that the

interests of different invention users will be somewhat disparate, such that the benefit which each party

will expect from a new widget container design will be different for each potential user. In these

circumstances, some users (the least benefitted by a new advance) will expect others (the most benefitted)

to pay more of the total administrative and incentive costs for development of the innovation (perhaps via

a contribution to these costs that is scaled to the projected benefit of the targeted advance to each user).

This type of differentiation of interests -- and the many related opportunities for disagreements and

frustration of contract formation and administration processes -- adds substantial further uncertainty and

dysfunctionality to contractual means for promoting innovation on behalf of multiple users.

E. Patents as Substitutes for Contract-Based Incentives In Agency Innovation

By solving some of these problems with contract-based agency processes, patent rights offer

streamlined means to promote innovation by multiple innovators on behalf of multiple potential

innovation users. Patent rights provide means for overcoming the problems of group action for both

principals and agents in innovation projects.

For principals (that is, potential invention users), patent rights create incentives for innovators to

organize and take risks on product development on behalf of groups of potential invention users. These

risk-taking innovators are often companies given the necessary resources and scale of activities needed to

conduct many modern forms of technological research. Corporate innovators serve as stand-ins for the

interests of multiple users, taking innovation risks cabined (at least in part) by confidence based on patent

rights that if the companies are successful in producing new products desired by users, the users will need

to come to the companies for the resulting products and that the costs of development and a further profit

compensating for risk taking will paid from product sales. The need for contracting in advance with

users is superseded by the patent-guaranteed promise of exclusive domains of commerce in a patented

advance. Corporate innovators can focus on risk taking and capital generation concerning technology

capabilities and risks of technological project failure without worrying about the further risks of free

riders and other compensation gaps upon production of technologically successful designs. Innovation

companies accept the risks of innovation development and avoid the need for costly group contract

formation. They also preclude free rider holdouts among invention users by demanding that all users pay

patent-influenced prices if they want access to patent protected products.

For agents (inventors), patent rights create value in their research results that they can not only

convey to others, but that varies in amount based on the full scope of benefits of successful inventions to

their complete set of users (at least in the geographies and time frames of patent control). This valuation

scaling reflects both the value of a patented advance to each user and the number of benefitted users. A

patented invention creating an average of $1000 of incremental value for each user over prior products

44

lacking the patented feature is more valuable than a patented invention with $100 of incremental value, all

other things being equal. And a patented invention with $1000 of incremental value to 1,000,000 likely

users is more valuable than a patented invention with $1000 of incremental value to only 1,000 likely

users, all other things being equal. Inventors can project these features of future value and target their

innovation efforts accordingly.

This projection of value will have differential effects on both the invention targeting and

invention perfection efforts of potential innovators. This value scaling will encourage inventors to

diagnose the needs of users to serve the largest individual needs possible and the largest numbers of

potential users as possible. It will also encourage them to shape their actual innovation efforts and

designs to serve the largest aggregate value to users. By thinking in this way, innovators are incentivized

by patent rights to identify and then pursue the aggregate interests of potential users of patented advances.

In this way, patent rights have the potential to create the equivalent of contract-based invention

incentives, taking into account both individual users benefits and overall user numbers, without the need

for costly contract formation and administration processes.

V. Patent Law Through an Agency Theory Lens

By examining the impact of patents in shaping agency relationships for innovation, we can

identify a number of previously unappreciated influences and functions of patent rights. Patent standards

and policies of several types deserve further study and possible reforms in light of their agency roles.

This section identifies some of the patent issues that may bear new study when seen from an agency

perspective. The series of patent law issues addressed here does not exhaust the range of patent questions

that may be profitably reevaluated via an agency lens, but does suggest the power and value of agency

evaluations of patents across a wide range of patent standards and practices.

A. The Outer Boundary of Patent Rights: Desirable Limits of Patentable Subject Matter

Under an Agency Model

1. When Should Patent Rights and Rewards Substitute for Contract Terms as

Innovation Incentives?

Patentable subject matter standards establish the outer boundaries of the patent system and

associated administrative processes.60

Types of innovations that do not constitute patentable subject are

simply outside the patent system and related rewards, controls, administrative processes, and enforcement

costs. Hence, patentable subject matter standards provide means to tailor the patent system to situations

where the special incentives of that system are needed, while excluding the system from situations where

the net costs of the patent system are high.

Viewed as a substitute for contractual processes in creating desirable agency relationships for

innovation, the proper scope of the patent system can be evaluated as a response to contract failure.

Under this view, patent rights are substitutes for contract rights in situations where contract processes are

likely to fail and consequently unlikely to form desirable agency relationships for pursuing innovation

projects. The desirable scope of the patent system – and the proper scope of patentable subject matter –

is dictated by likely contract failure. Patent rights are desirable where they are needed as a contract

substitute in forming agency relationships for innovation. Conversely, where contracting for innovation

60

Advances that meet other patent law standards are only proper targets of patent protections if they entail

patentable subject matters. See 35 U.S.C. § 101. Thus, the subject matter tests within patent law are

effectively “gate keeper” standards determining where the patent system will have any traction and effect.

45

is likely to be effective in serving the needs of all likely users of an innovation, then the administrative

costs (and potential intimidating effects) of the patent system probably serve no incremental purposes and

should probably be avoided. This section develops an account of patentable subject matter standards

from this agency perspective, showing how agency analyses establish a normative case for specific

patentable subject matter standards (or at least portions of such standards).

Viewing patent rights as useful substitutes for contract-based agency innovation processes, these

substitutes may be particularly valuable in at least three circumstances. First, patent rights may be needed

as contract substitutes for innovations that are likely to be of interest to large numbers of potential users.

Second, patent rights may be needed as contract substitutes for innovations that are of interest to only a

few users, but that are within the capabilities of only a narrow number of innovators who are difficult to

identify and contractually influence. Third, patent rights may be needed as contract substitutes for

innovations depend on new technologies with functional attributes that are largely unknown to potential

users such that these users will be unlikely to contract with specialists in those technologies as means to

serve the users’ functional needs. This section considers each of these settings where patent rights may be

of particular value in adding meaningful inducements for desirable innovation.

2. Innovations Foreseeably Capable of Productive Use By Numerous Parties

Patent rights will tend to be useful substitutes for contract-based innovation incentives where

innovations appear likely to serve numerous users. Contract substitutes will be needed in this context

because innovation contracts incentivizing agents to advance the full aggregate interests of the relevant

users will be difficult to form due to the large number of relevant users. Where an innovation is likely to

be broadly useful to numerous users, the binding together of the relevant user group thorough contracting

processes will be particularly burdensome due to the difficulty of identifying and negotiating contract

terms with most or all of the substantial number of parties involved. Without full participation of the

interested parties in innovation contracts providing promised rewards to innovation producers, such

contracts will only reflect the lesser interests and incentives provided by the subset of users within the

contractual arrangements. These incentives will be suboptimal because they are not based on the full set

of interests and potential invention benefits for all the invention users. The difficulty in forming contracts

among the user set will ensure that only some of the relevant users are involved, which will in turn ensure

that only some of the relevant incentives for innovation are provided to agents considering and pursuing

innovation projects.

Rather than relying on contractual processes, rewards created by the enforcement of patent rights

can establish potential rewards that reflect the aggregate utility of an advance to the full set of its potential

users. The full set of users can be charged for the resulting innovation because all users will be

precluded by patent rights from uncompensated access to the inventions. The charges paid by this full set

of users can pay for innovation incentives that reflect the full number and extent of user interests served

by successful inventions. These sorts of desirable invention incentives for agents pursuing innovation can

be created through patent rights while avoiding the contract formation costs and free rider effects

associated with private contracting processes aimed at encouraging innovative efforts by agents.61

61

I have previously argued that these considerations regarding the limits of contractually-initiated innovation

indicate that certain advances in computer implemented information processing methods should be

recognized as patentable subject matter. In particular, my analysis suggests that information processing

advances that are capable of description in a manner that can be transferred to multiple users and that are of

use to a substantial number of users should be treated as patentable subject matter. See Richard Gruner,

Intangible Inventions: Patentable Subject Matter for an Information Age, 35 Loyola L.A. L. Rev. 355, 449-

450 (2002).

46

3. Innovations By Hard to Find Innovators

A second class of innovations that are also poor targets for contract-based innovation (and

correspondingly good targets for patent-based substitutes) includes innovations capable of production

only by innovators who will be hard for users to identify in advance of innovation efforts. The difficulty

in identifying candidates for producing particular types of innovations may stem from the rarity of

individuals having the necessary skills or knowledge to produce the needed innovation, the lack of public

revelation of the full extent of the innovators’ knowledge or skills, the unfamiliarity of the innovators’

work to the relevant users because that work is performed in a field having little contact with the user

group, or the lack of special features distinguishing innovators from large numbers of similarly situated

parties such that the innovators are lost in the crowd of similar parties. In these and other circumstances,

difficulties in finding relevant innovators will impede the formation contract-based relationships for

innovation. If interested users cannot find the relevant agents capable of serving the users’ interests, they

cannot form contracts incentivizing innovation by the agents. Hence, difficulties in identifying innovators

establish weaknesses in contract-based processes for innovation, which creates corresponding needs for

patent-based incentives.

Difficulties in identifying relevant innovators may be quite common and lead to frequent gaps in

contracting processes. Search costs for relevant innovators may be large and neither set of interested

parties -- the potential users of a new advance and the parties with sufficient skill to develop the advance -

- may have a sufficient stake in the outcome to invest the search costs needed to find each other. Search

costs may be particularly high in the early stages of the development of a new technology or engineering

approach because the parties with knowledge and skills regarding the technology or design approach will

be poorly known. Also, in the early development of a new technology or engineering approach, the

individuals who have exceptional knowledge in the relevant discipline may be few in number. Even if a

few of these parties are publicly known in the early development of a technology, they may also be

contractually obligated to projects in the relevant field, leaving only other, harder to find parties as

candidates for additional innovation design projects.

For example, if computer chip makers have difficulty with a particular chip cleaning problem,

these specialized manufacturers may have little reason to look for a solution in the food canning industry

where a similar cleaning problem had recently been solved and even less ability to identify the leading

experts in the food canning solution. Even if the chip makers were aware of some parallel cleaning

concerns and technology applications in the chip manufacturing and can cleaning contexts, persons in the

chip making field might have little way to monitor when persons in the can cleaning field reached insights

into cleaning processes that might be used in chip making. The chip manufacturers might also have few

insights into who to contact within the food canning industry for design help in developing a new chip

cleaning technology. Hence, it is unlikely that the chip makers would pre-identify and establish

innovation contracts to incentivize parties in canning company to act as agents of the chip manufacturing

concerns in the development (or modification) of can cleaning technologies in ways that were peculiarly

interesting and beneficial to the chip manufacturers.

The barriers to innovation in these settings stem from transaction costs limiting successful agent

identification and related contract formation. Anticipating what type of innovator to look for, identifying

an innovator with sufficient ability to produce a cross-field advance, and then concluding an innovation

contract with that innovator on terms agreeable to both the innovator and the relevant innovation user are

all steps with significant transactional difficulties and costs. Companies seeking innovation will not want

to add these costs to the costs of actually completing innovation projects and will tend to forego this type

of multi-field search for innovators in favor of innovation attempts within their company or industry. As

a consequence, it is unlikely that the parties will form incentivizing contracts linking the interests and

perspectives of innovators in the canning industry to those in the chip manufacturing. Without this type

47

of contract (or some other source of cross-industry incentives) parties in the canning company will retain

their limited perspective on the potential value of their new technology and will not seek to extend or

clarify the value of that technology as for applications in the chip manufacturing domain. Experts in the

canning industry will limit their perspective and technology-development project scope because there are

no contractual incentives to take a broader view and pursue broader goals. They have not been

incentivized to serve the full set of potential users of their advance.

Patent rights offer means to overcome this incentive gap. Non-contractual rights and sources of

invention rewards arising from patent rights encourage innovators to pursue value in new fields and

formulate their advances as broadly as patent rights and associated rewards will sweep. This perspective

setting feature of patent rights operates without the need for pre-existing identification of the relevant

innovators and innovation users. Nor does it require the formation of contracts between the parties.

Rather, the promise of patent rights encourages innovators to look broadly for value and to consider the

interests of as many users as possible from the outset of an innovation project, with the actual

identification and charging of innovation users deferred to a later point when an actual advance is in hand

and patent rights for that advance can be enforced.

Patent rights encourage innovators and innovation users to find each other and establish

associated agency payoff arrangements. Patent rights do this through two processes.

First, by encouraging parties who are knowledgeable about new technologies to adopt broad

perspectives in looking for applications and commercial opportunities for the technologies, patent rights

encourage parties working on the design, refinement, and implementation of advances that serve the

broadest set of potential users and largest aggregate value to users (since this will tend to produce the

largest patent-mediated rewards for innovators). This perspective setting encouraging attention to

aggregate user needs will influence both the designs of new applications of technologies and the size of

efforts by innovators to realize applications. Application design projects with the potential to serve large

user sets and achieve significant aggregate value for those users will justify the largest application design

and development efforts. Patent rights exert desirable influences over the design and refinement of

technology applications because design projects are incentivized from the outset towards tailoring in light

of the viewpoints and interests of invention users. Agency relationships for innovation are established

from the bottom up by patent incentives because the potential for patent enforcement and related

commercial returns encourages innovators to look at their work with the eyes and design goals of all the

potential users of their inventions.

Second, patents (or, at even earlier stages, the publication of patent applications) reveal new

technologies to the public in relatively completely described forms, thereby providing means for potential

users of advances based on the new technologies to identify the new technologies as potential bases for

new applications. This will help other parties in the same field as the inventor to understand the new

advance (and either work with it under a license from the patent holder or work around it through the

creation of new designs that do not include the patented advance and, hence, do not infringe the related

patent). The public description of the new technology will also help users in fields not initially

considered by the original innovators to understand the patented advance and to consider whether the

advance may also have value in these further users’ fields. Through this response to publicity for the new

technology (or even its availability in computer searching processes), a patented technology transfer

process can be driven by the needs and actions of innovation users (or technologists acting on behalf of

those users). Principals (that is, potential innovation users) in diverse fields can seek out agents who have

already developed the germ of a potentially useful advance and license the patent rights to use the

advance in additional fields and various products or services far beyond those contemplated by the

original technology innovator. The original innovators will have strong incentives to cooperate with

these efforts by granting the licenses and aiding parties in the additional fields since this will expand the

48

applications and potential patent-mediated rewards available to the innovators. The initial innovators will

have patent-generated economic motivation to serve as agents of the innovation users in the additional

fields because this will expand patent rewards. Here, patent rights and associated downstream licenses

establish links after the initial development of a patented invention that cause the original inventors to

serve thereafter as agents of innovation users in aiding the development and commercialization of

diversely-focused products and services incorporating the patented advance.

4. Innovations with Initially Concealed Value

Information gaps hindering the perception of innovation value by both innovators and potential

innovation users may also impeded the formation of contracts incentivizing innovation. If potential users

are unable to see any advantages from a particular type of technology (and innovators are unable to

convince the users of this value), the users are unlikely to bear the costs of incentive contracts

encouraging agents to develop the technology. Without evidence of the value of a new technology, users

funding agents to develop the technology would feel that they were throwing their money away on

apparently unproductive tasks. Hence, factors concealing the value of a projected new technology are

also likely to stop the formation of agency relationships promoting the development of that technology.

The new and often highly unfamiliar nature of a new technology may easily conceal its value.

Where – as will be the case with many of the non-obvious advances qualifying for patent rights -- an

innovation comes from outside the normal technology development channels in an industry, the parties

producing the innovation may have far more understanding of the operation and functional characteristics

of the innovation than any other parties in the same field. Potential users of products and services

incorporating the new technology – who may not even understand technologies underlying long-standing

products and services in their field – may frequently lack basic information needed to gauge the

functional impacts and value of the new technology. Their lack of information about innovation value

may cause users to forgo contracts for further development of the new technology even though such

contracts would provide gains to the users and be attractive to them under conditions of perfect

information.

Patents can provide incentives for innovators to overcome these information gaps and

demonstrate to users the value of patented advances. Under the incentives created by patent rights,

innovators can count on patent-mediated rewards which will tend to be in amounts that track the

perceived value in the hands of users. The more users see and embrace the value of a new advance and

seek products and services incorporating the advance, the more a patent holder stands to gain in patent-

mediated sales prices or licensing royalties. This is the case because patent rights will force consumers

who which to gain access to the new technology and realize its potential value will only be able to get the

technology from sources authorized by the patent holder. The patent holder will be confident of gains

from user understanding of the value of a new technology and will take the steps needed to expand that

understanding. In this way, the rewards promised by patent rights will make patent holders the promoters

of their own inventions, with an emphasis on establishing and publicizing the practical value of the

invention to as many parties as possible. Patent rights will help to encourage innovators to bridge the

information gaps that would otherwise stall the formation of incentive contracts for innovation and

impede the development of potentially valuable new technologies.

Patent rights incentivize innovators to conduct searches into uses for a new technology without

the need for the formation of contracts before those uses are established in the eyes of potential users.

Patent rights ensure that innovating agents are offered rewards that relate to the full ultimate value of

advances to users (at least during the term of the patent rights), but rewards do not depend on any

particular user or users appreciating the value of the advances at the stage when agency contracts seeking

innovation would need to be concluded. Rather, the patent system defers the point of fee settlement for

49

an innovating agent -- that is, the determination of the proper amount of payment to the agent for a

successful innovation -- to a point when innovation users have much more information about the impact

of the innovation and the full benefit that should be the basis for the fee to the innovator. Patent rights

encourage innovators to rely on their own superior information about a technology in its early stages of

development, not only about the engineering aspects and chances for technical success of their projected

innovations, but also their conclusions about the probable value of innovations to the full range of users

who are ultimately likely to perceive and realize benefits from the innovations.

B. Timing and Content of Patent Application Disclosures

Another feature of patent law that may bear reexamination from an agency standpoint is proper

scope and timing of pre-issuance patent disclosures. Through an agency lens, these disclosures are means

to either enhance the work of agents promoting innovation or to cut off duplicative and wasteful efforts of

multiple agents who are working in parallel on similar innovation projects. Disclosures through the

publication of patent applications can cut off potentially duplicative and wasteful parallel innovation

efforts by multiple agents who are seeking (or may be about to start seeking) to fill a single type of user

need and thereby receive a corresponding reward. By modeling parallel innovation efforts as instances of

competition between agents seeking a single payoff from a group of principals, disclosures in patent

applications can be treated as signals from one disclosing competitor to others about the state and

direction of the disclosing party’s work. Such a signal can cause other innovators pursuing parallel

innovation efforts to make one of several responsive actions: 1) ceasing their similar efforts (because the

efforts will duplicate those already completed by the party filing the patent application), 2) redirecting

their efforts to ensure that they are pursuing a substantially different engineering approach which will not

compete with the already developed and disclosed approach (and which may be able to qualify for its own

patent protections in addition to whatever patent rights result from the published patent application), or 3)

gaining new confidence that they are already on a second, fundamentally track from the advance in the

published patent application, thereby reducing their perceived risk since at least one potential competitor

(the party revealed via the published application to have taken a different track) will now no longer be

seen as a viable threat for earlier and patent-blocking development of the second technology.

All of these responses will reduce waste due to unproductive efforts and tend to increase the

range of new technologies considered in the relevant field. The researchers involved -- those receiving a

signal or "technological status report" on the efforts of a competing innovator via a published patent

application -- will gain the most if they receive this sort of signal as early as possible. The earlier that it is

apparent that another party has succeeded in producing some new technological development, the more

duplicative research that can be avoided and the earlier and more completely fears of superseding

innovation development can be prevented. Hence, an agency model of patent law suggests that the

earliest possible public disclosures of patentable advances is desirable, not for the consumer protection

reasons related to "submarine patents" that are often cited as the basis for pre-issuance publication of

patent applications, but rather because the earliest possible disclosures promote desirable signaling to

competing innovators. This signaling at the earliest possible points also supports prompt rechanneling

and reassurance of additional innovation by competitors of the disclosing innovators. Early signaling

regarding the nature and success of innovation attempts should also support diversification of the

innovation approaches carried out through agency relationships aimed at fulfillment of a single type of

user need.

The contents of invention disclosures that are desirable in a published patent application can also

be evaluated from an agency perspective by considering the impact of various types of disclosures on the

efforts of researchers competing with the disclosing party. Present patent laws require disclosures in

patent applications of information needed to recreate or "enable" use of the disclosed technology. Patent

applicants are required to provide other engineers in the relevant field of an advance with enough

50

information to recreate and implement the patented invention.62

This type of disclosure aids in

transferring patented advances to the public at the end of a patent term (when all parties should be able to

use the advance without constraint). However, an enabling disclosure alone may not maximize the

potential impact of a published patent application in cutting off wasteful and duplicative research. More

effective disclosures aimed at avoiding waste in competing agency processes for innovation might better

serve the public.

If required disclosures in a published patent application are viewed as means to prevent wasteful

duplication of innovation efforts, a different range of disclosures than is currently required may be

advisable. Disclosure requirements might be reshaped to better signal fruitless lines of inquiry to other

competing innovators. In addition to providing enabling disclosures for the benefit of potential invention

users, applicants might be required to describe a complete history of their innovation efforts leading to the

conception and reduction to practice of their claimed invention. Such as research history would signal to

other innovators both the applicant’s failures and partial successes in innovation efforts. Such helpful

disclosures about engineering progress and "dry holes", given to the engineering community and the

public in exchange for the patent rights being sought in a patent application, might be desirable means to

increase the impact of published patent applications in preventing duplicative agency processes for

innovation and in directing subsequent innovation efforts by agents towards still promising yet

unexplored technological directions.

C. Interpreting Non-Obviousness Standards

Unites States patent laws limit patents to new advances that are non-obvious in light of pre-

existing, publicly available products, services, and technical knowledge.63

Non-obvious inventions

meeting standards are unlikely to be made by the bulk of innovators in a given field. A non-obvious

advance is one that a well-informed specialist having average skill in the relevant field would be unlikely

to produce. Typically, these non-obvious and patentable advances will be within the range of only well-

informed parties of greater-than-average skill in the relevant technological field. However, the number

of innovators having greater skills or knowledge than the average innovator may be relatively few.

Hence, non-obviousness tests effectively restrict patents to advances that are only within the capabilities

of a relatively rare set of innovators in a field – those capable of advances requiring greater-than-average

skills or knowledge.

An agency view of patent law provides an explanation of why this type of special focus on

inventions by rare innovators is important. Where only relatively few innovators are potential candidates

to produce particular types of advances, contracting processes to find and engage those few innovators in

agency relationships to produce these advances may be difficult to carry out. Contract-based incentives

for innovation (at least with respect to non-obvious forms of advances) will be correspondingly rare or, if

present at all, diminished in value by the high cost of the contracting processes. A non-contractual

incentive setting mechanism such as that resulting from patent rights is particularly valuable in these sorts

of settings where contracting is difficult due to the rarity of the capable innovators. Non-obviousness

tests focus patent rights on the settings involving rare innovators where non-contractual innovation

incentives are specially needed. These same standards withhold patents and associated patent

enforcement costs for advance that are within the capabilities of numerous innovators and for which

contractual arrangements motivating related innovation efforts are more likely to be easily concluded.

62

See 35 U.S.C. § 112.

63 See 35 U.S.C. § 103.

51

Patent incentives will be superior to contract-based incentives for activating development of non-

obvious advances not only because patents avoid the need for contract formation, but also because the

patents help to ensure that the full range of innovation users are compelled to contribute to innovation

rewards through patent-mediated sales prices and licensing royalties. Patent enforcement compelling all

users of patented advances to contribute produces greater innovation incentives than might be paid by a

few contractually-bound users sponsoring innovation. Patents tend to ensure that the full range of

innovation users can be called upon to pay for access to the advances during the life of particular patents.

This will tend to ensure that perceived patent rewards increase with the full value of advances to all

potential users. Promised rewards at these high levels will be important for non-obvious advances

because the rarity of the parties capable of producing these advances means that they will often have

many possible choices about activities to pursue and many competing offers of compensation luring them

from the production of new inventions. To trump these competing callings and induce a party with rare

skills or knowledge to address an important advance, the party may need to be offered highly substantial

rewards. Patent-dictated fees based on the payment capabilities and interests of the full set of potential

users of the resulting advances are desirable means to increase promised innovation rewards to the

necessary levels.

D. Achieving Blended Rewards for Patent-Influenced Innovation

Agency interpretations also allow us to consider the impacts on innovation motivation of complex

incentive schemes offered to innovators. The advantages of separately allocating innovation risks and

incentives can be expressly modeled through an agency approach. In addition to examining the mixture

of efforts-based and results-based incentives that can be offered to innovators under a patent-framed

reward system, we can evaluate the advantages of establishing separate risk allocation arrangements at

several levels of multi-level agency processes for innovation. In particular, using agency analyses we can

appreciate the advantages of the risk allocations achieved by involving corporations (who are relatively

risk neutral actors) as the primary agents searching for new innovations serving users, with the further

involvement in direct innovation activities at a lower sub-agent level of individual employee-researchers

(who are generally risk averse actors).

1. Multi-Level Agency Processes Implemented via Corporate Innovation

The promise of patent rights and returns for corporations whose employees produce useful

advances may both encourage risk neutral innovation efforts by corporate entities and accommodate the

risk aversion of the employees who innovate on behalf of the corporations. Corporations will typically

have much larger resource pools to allocate to research than individual innovators, leading the

corporations to be much less concerned about resource exhaustion as a result of failed projects than would

individuals pursuing the same projects. The resulting risk neutrality of the corporate actors will allow

them to support innovation projects with an even hand – that is, to see projects as cost-effective and

desirable or undesirable with a neutral view of the risks of failure and return associated with the projects.

Risk neutrality will tend to aid in the alignment of corporate commercial interests with measures of

aggregate value of advances to innovation users. As such, risk neutral corporations may be better project

initiators than relatively more fearful and risk averse individual innovators. Risk averse corporations will

tend to support and initiate some innovation projects that are in the public interest (given their projected

costs, benefits, and risks of failure), but which are not attractive to risk averse individuals who must worry

about the personally devastating consequences of innovation project failure and the individuals’ loss of

income and basic assets.

In addition, corporate innovators may be better at targeting the initiation of innovation projects

because they are working with different targeting information than individual innovators. Based on both

their accumulated information about customers and their further experience with similar innovation

52

projects, companies can sometimes use their insights to see both more of the advantages (in user value

and commercial potential) and disadvantages (in technological failure and innovation process challenges)

of potential innovation projects. Because they are able to approach innovation projects with relative risk

neutrality and accumulated information about innovation processes and consumer needs and desires,

corporations acting as innovation agents -- or at least as primary agents much like general contractors in

construction projects – can play important roles in advancing innovation projects. Acting at their full

potential, corporate innovators can be superior to individual actors in selecting innovation projects on

behalf of potential users, as well as in generating the financing and coordinated resources needed for large

scale innovation projects.

Corporations may serve several valuable functions as primary agents standing between

innovation users and individual innovators. Past experience with users can help companies to build up

information about users’ likely needs and reactions to future product designs. This information will aid in

defining innovation goals for projects by agents working in favor of the interests of innovation users.

Furthermore, parties within corporate marketing departments may be able to estimate the size of various

product sales opportunities (including estimates of both sales volumes and potential product pricing) to

combine spectrums of product demand to identify weighted estimates of the most strongly desired

products. These estimates can be translated into engineering priorities for product or technology

development and funding personnel. Along parallel lines, product manufacturing and distribution

specialists may be able to contribute estimates of product production and sales difficulties that will

diminish the desirability of projected types of products, thereby shifting product development priorities

away from such products.

In these functions, corporate specialists other than technology specialists are acting as user agents

to organize and prioritize the actions of their companies along the lines that will best serve perceived

needs of product users. The specialists both project user needs and estimate the abilities of their company

to meet those needs. From these estimates, they can generate some rough corporate priorities for future

technology development. These corporate activities constitute agency actions that focus and advance

technology development steps on behalf of potential customers. The result is technology development

that more closely tailors the activities of the companies as innovators to the agency interests of the

targeted technology users. Patents enhance these agency processes by framing the consumer demand and

related commercial potential for technologies considered by companies, providing reassurances that the

companies can look to the full extent of a patent-controlled consumer demand for a new technology as a

payoff for developing that technology. Patent rights (and the exclusive commercial opportunities

maintainable by enforcement of the rights) will reassure company executives that, if their company can

produce products with distinctively new, patentable features of value, the company can count on gaining

the full commercial return on those new features for the life of the applicable patents.

2. Blended Incentives to Accommodate Individual Risk Aversion

Agency analyses also suggest how individual-level incentives should be structured once a

technology development project is initiated. Some mixture of rewards for technology development

efforts (involving relatively low risks of non-payment to the individual innovators) and further rewards

for technology development results (involving higher risks of technology or project failure) may be

needed to accommodate the risk aversion of individual innovators. In practical terms, this means that the

risk aversion of individual innovators may be best accommodated by incentivizing them to pursue

innovation via a promised mix of salary payments and further bonus or profit sharing payments upon

successful project completion. Risk aversion may also make individual innovators concerned about their

own limited resources to be poor decision makers about targeting innovation since they will tend to shy

away from some risky projects that are too risky for the risk averse individual innovators yet

commercially advantageous (and publically beneficial) if looked at from a risk neutral viewpoint.

53

By splitting corporate, risk-neutral innovation targeting and invention returns from individual,

risk averse incentives for pursuing innovation tasks, both the targeting and motivational problems

stemming from the risk aversion of many individual innovators can be solved. The compensation and

innovation incentives provided to company employees as sub-agents of innovation users (with their

corporate employers as the primary agents) can be structured in many ways, with the dual aims of

matching the risk preferences of the employees and encouraging them to undertake the actions desired by

their companies. Matching the risk preferences of employees will usually entail accommodating the risk

aversion of these individuals regarding the potential loss of all or most of their income due to a failed

innovation project. Combinations of incentives for innovation efforts and results can be built into

employee compensation schemes without exposing the employees to the full risks of innovation failure.

Specifically, a combination of rewards for efforts (such as low-risk salary rewards granted to employees

upon completion of assigned tasks) and rewards for results (such as higher-risk bonuses for production of

patented inventions or even larger but riskier bonuses based on the commercial success of inventions)

may be the best way to incentivize innovation efforts while accommodating individual risk aversion

tastes.

Companies standing between product users and individual innovators provide a useful risk

adjustment buffer by both taking innovation risks in a risk neutral manner while passing on risk averse

incentives to individual employees undertaking innovation projects. Corporate innovators attempt to

translate the demand of potential product users into product development goals and projects with related

innovation incentives for individual innovators acting as technology and product developers. A corporate

employer can create employee incentives representing and reflecting the commercial interests of product

users to the company’s employees. The company can achieve this by providing compensation to

innovators that encourages innovative efforts with a focus and to a level that is efficient in light of the

projected utility gains to the full set of users of the targeted technologies and products. At the same time,

by providing blended combinations of efforts-based and results-based incentive compensation, corporate

employer can present blended compensation terms to innovators that offset the innovators' risk aversion

and encourage optimal efforts by the individuals to pursue advances serving numerous users.

These sorts of blended incentives to individual innovators can be provided through various types

of corporate employment compensation, attracting many individual to innovation projects who might

otherwise shy away from such efforts due to risk aversion. Individual innovators who choose to work for

corporations or other organizations typically give up large project failure risks in favor of relatively stable

income and lifestyles. If they choose to work for corporations as employees, innovators must typically

give up (that is, assign) their patent rights to their corporate employers. In exchange, the employees

receive a salary for their innovation services. However, many corporations also provide employees with

results-based bonuses that turn on innovation success (sometimes scaled to the revenues generated by a

successful invention). In both the amount of their flat compensation for innovation efforts and in the

degree of results-oriented bonus compensation added to the flat fee, companies relying on projected

patent rights can provide compensation levels to employee-innovators that vary with the full commercial

value of anticipated products across broad user groups. The scope of exclusivity promised by patents

ensures that the full commercial value of a patented advance can be captured and parallel, full-value

influenced incentives passed on to employees completing innovation projects.

3. Blended Innovation Incentives in Universities

Universities can also blend compensation schemes for academic innovators to help ensure that

these innovator focus on the needs of innovation users and the aggregate value to users of potential

advances. In academic settings, however, the reward types (and hence the reward mixes) that will be the

best motivators will vary from the packages used in corporate settings. In particular, the best efforts-

54

based rewards for academics may be somewhat different than those for corporate employees. In addition

to providing salaries to academics for research work related in part to the completion of patentable

inventions, universities can provide academic innovators with further non-monetary compensation by

supplying stimulating intellectual environments and reputational rewards corresponding to academic

positions and activities. On top of these rewards for innovation efforts, universities can provide

academics with further results-based incentives such as portion of the royalties resulting from patentable

inventions. More substantial results-based rewards to individual academics may flow from university

policies allowing flexible leaves of absence to academic innovators which permit the innovators to join

startup companies and share in the commercial value of their discoveries as co-owners of these firms.

Through these types of opportunities, academics are given the chance by their university employers to

realize gains from the work of the individual academics in amounts that are scaled to the full user

functionality and commercial potential of the academics’ patented inventions. Working in a university

environment creates incentives for innovators to serve as agents of product users, with the features of the

incentives tailored to the preferences of academics through combinations of efforts-based compensation

(through salary and environmental perks) and additional compensation tied to commercially results of

innovation projects. Such a combination of blended incentives permits universities to match the

particular risk and incentive preferences of academic innovators and to encourage innovation by these

parties aimed at producing inventions of use and value to broad segments of the public.

VII. Conclusion

By examining patent rights and related innovation processes from an agency perspective, the role

of these rights in encouraging innovation efforts on behalf of the public can better be understood. This

understanding can expand our descriptive understanding of patent impacts in many ways. In particular,

patent analyses within an agency framework can directly illuminate the risk allocations and informational

disparities surrounding patent-influenced innovative processes.

Agency analyses can also expand our normative evaluations of the patent system. The normative

implications of patent rights can be modeled by viewing these rights as sometimes valuable alternatives to

contract-based rewards in establishing and incentivizing agency processes for innovation and for

adjusting innovation risks and informational disparities between innovation producers and users. Patent

rights -- in combination with related institutional contracting arrangements such as employment contracts

in corporate and university settings -- can create innovation incentives for potential innovators that tailor

incentives to the aggregate user value and demand for new advances. In some settings, this tailoring to

user demand and value assessments can be achieved more effectively via patent rights than direct

contracting with users will allow. Patent system scope should extend to those contexts where patent-

based incentives for innovation appear more likely to reflect full user demand than comparable

contracting processes. Consequently, the scope of our patent system should depend, at least in part, on

assessments of where private contracting for needed innovations is likely to be ineffective and where

further, patent-enhanced rewards can therefore serve a valuable function as contract substitutes in forming

and administering agency processes for innovation.

Where contracting processes are particularly unlikely to succeed in producing innovations -- as,

for example, with non-obvious innovations that can be produced by relatively few and difficult to identify

researchers -- patent incentives may be the only practical way to encourage innovators to adopt public-

oriented perspectives on innovation goals along particularly risky paths of technology advancement. By

focusing strong, aggregate demand influenced patent rewards on the achievement of non-obvious

technological advances, patents encourage potential innovators to prioritize and pursue non-obvious

technological advances in light of how these advances that will serve aggregate pubic needs. These

strong and substantial rewards encourage innovators with unusual skills and knowledge to significantly

depart from prior learning in intellectually risky manners to develop previously non-obvious advances,

55

and, with their occasional successes, to materially broaden the range of useful technological tools applied

for public benefit. In this way, patent rights and the public-serving agency processes they further are

critically important features of modern innovation, providing essential and highly valuable links between

private innovation incentives and public benefits from expanding types of successful inventions.